Studies On Product Design Using Ergonomic Considerations: Doctor of Philosophy

Studies on Product Design using
Ergonomic Considerations
A THESIS SUBMITTED IN FULFILLMENT OF
THE REQUIREMENT FOR THE AWARD OF THE DEGREE
OF
DOCTOR OF PHILOSOPHY
IN
DEPARTMENT OF INDUSTRIAL DESIGN
BY
PRAGYAN PARAMITA MOHANTY

(ROLL NO.511ID102)
NATIONAL INSTITUTE OF TECHNOLOGY

ROURKELA - 769008, INDIA
NOVEMBER – 2014
NATIONAL INSTITUTE OF TECHNOLOGY
ROURKELA-769008, ODISHA, INDIA
CERTIFICATE OF APPROVAL
This is to certify that the thesis entitled STUDIES ON PRODUCT DESIGN
USING ERGONOMIC CONSIDERATIONS submitted by Pragyan Paramita
Mohanty has been carried out under my supervision in fulfillment of the
requirement for the award of the degree of Doctor of Philosophy in
Industrial Design at National Institute of Technology Rourkela and this
work has not been submitted to any university/institute before for any
academic degree/diploma.
Prof. Siba Sankar Mahapatra

Professor
Department of Industrial Design

National Institute of Technology, Rourkela-769008, INDIA
This thesis is dedicated to lord Gayatri, my son Abhigyan
and all who have inspired me.
ACKNOWLEDGEMENT
While bringing out this thesis to its final form, I came across a great number of
individuals whose selfless dedication and contribution in various ways that helped me
to execute this research work and therefore, they deserve special thanks. I am
greatly indebted to them.
First and foremost, I would like to express my deep sense of gratitude and
indebtedness to my supervisor Prof. Siba Sankar Mahapatra, Professor,
Department of Mechanical Engineering, National Institute of Technology Rourkela, for
his esteemed guidance, invaluable encouragement, moral support, free to act on my
ideas and scholarly inputs from early stage of research work that instill confidence in
me during the research work. Above all, his priceless and meticulous supervision at
each and every phase of work has been the cradle of illumination for me. This thesis
could not have been completed without his accordant suggestions, motivation,
constant encouragement and crucial contribution, which have enriched value of my
thesis. He had always been there to route me through all pros and cons.
Besides my supervisor, I would like to thank the rest of my doctoral scrutiny
committee (DSC) members: Chairman (DSC Member), Prof. B. B. Biswal, Head,
Department of Industrial Design, Prof. B. D. Subudhi, Professor, Department of
Electrical Engineering, Prof. S. Datta, Assistant Professor, Department of
Mechanical Engineering and Prof. H. B. Sahu, Associate Professor, Department of
Mining Engineering for their encouragement and insightful comments.
I am highly grateful to Prof. S. K. Sarangi, Director, National Institute of
Technology (NIT), Rourkela and Prof. B. B. Biswal, Head, Department of Industrial
Design, Prof. B. Majhi, Dean (Academic Affairs), National Institute of Technology,
Rourkela for the academic support and the facilities provided to carry out the
research work at the Institute.
I also express my thankfulness to the faculty and staff members of the
Department of Mechanical Engineering, Department of Industrial Design for their
continuous encouragement and suggestions. Among them, Sri P. K. Pal, Sri J. K.
Sahu, deserves special thanks for their kind cooperation in non-academic matters
during the research work. Besides this I will certainly carry the fond memories of the
company of Research Scholars at NIT Rourkela for exchange of ideas and supports.
I am obliged to Sanjita Jaipuria, Bijaya Bijeta Nayak, Swagatika Mishra, Mr.
Swayam Bikash Mishra, Mr. Panchanan Jha, Mr. Chinmay Prasad Mohanty,
Mr.Suman Chatterjee, Mr. Kumar Abhishek and Mr. Chitrasen Samantra for their
i
support and co-operation that is difficult to express in words. The time spent with
them will remain in my memory for years to come.
I owe a lot to my parent, Mrs. Bimla Mohanty (Mother), who encouraged and
helped me at every stage of my personal and academic life, and longed to see this
achievement come true. I deeply miss my father Late Prof. Nirmal Chandra
Mohanty, who is not with me to share this joy.
I am very much indebted to my husband Mr. Abhay Sanatan Satpathy and son
Abhigyan for their understanding, patience, co-operation and support throughout the
course of my doctoral dissertation and I am highly thankful to them for bearing the
inconvenience of stay away from me.
I also feel pleased to strongly acknowledge the support received from my brother
Asit Mohanty and sister Sthitapragyan Mohanty in every possible way to see the
completion of this doctoral work.
Above all, I owe it to Almighty Lord Gayatri for granting me the wisdom, health
and strength to undertake this research task and enabling me to its completion.
PRAGYAN PARAMITA MOHANTY
ii
Abstract
Embedding ergonomic consideration into product/machine/equipment/component
design as well as work environment taking into account both psychological and
physical needs of user helps to enhance user efficiency, satisfaction and productivity.
It is vital to find best design elements to visualize the product which possesses the
characteristics not only to satisfy the users but also reduces fatigue and injury during
prolonged use. Although subjective and objective product characteristics are
important during product design, user comfort becomes a vital factor that can be
quantified by the analysis on continuous physical interaction between product and
user. Beside above influential factors, ergonomic design of product also considers
cognitive and behavioral information during the design stage with a view to improve
the comfort level of the user and aesthetic look of the product.
To address above issues, an integrated approach using statistical and artificial
intelligence techniques has been proposed in this thesis to effectively handle
subjective and objectives characteristics during design phase. The statistical method
is used to assess various user requirements and their significance whereas artificial
intelligence method determines the relationship between user requirements and
product characteristics. Since most of the psychological needs of users are difficult to
express quantitatively, combined approach of statistical and artificial intelligence
method can handle the subjectivity and uncertainty in an effective manner. The
approach has been demonstrated with the help of design of office chair. Keeping
view with the physical interaction between human soft tissue and product as a
measuring factor of comfort sensation in an office environment, a numerical analysis
of human soft tissue-chair seat model has been introduced into current work. In order
to evaluate superior ergonomically designed product (office chair), suitable multi-
attribute decision making (MADM) approach based on few important features has
been chosen to address the usability of product improving satisfaction level of
customer. The study also analyses a kinematic model of human upper arm extremity
to diagnose comfort arm posture that allows the operator to have a comfort work
zone within which possible postures can be accepted.
The integrated approach of statistical and artificial intelligence techniques
produces an office chair that satisfies most of comparable design elements of Bureau
of Indian Standard (BIS). Finite element (FE) model of the human soft tissue
(buttock)-seat predicts maximum stress in human soft tissue (at ischial tuberocity) on
prolonged sitting in an office environment. By the help of a detailed and realistic two
dimensional geometric description, the analysis provides insight into the problem and
iii
finds the ways to reduce the stress on bony prominence causing cell death of muscle
tissue and avoids suffering from pressure sore. The analysis also shows the effect of
postural changes on maximum stress beneath ischial tuberocity. However, a large
number of products are available in the market place possessing a wide range of
features to address the ergonomic considerations. In this regard, several multi-
attribute decision making (MADM) methods have been attempted considering both
subjective and objective weights for qualitative and quantitative design attributes for
selecting a suitable alternate (chair). In order to synergize the capability of an
operator within workstation, a comfort work zone has been generated with a
kinematic model of human arm. The model predicts an isocomfort posture of human
upper extremity to enhance the operator performance within a workplace. Model
efficiency has been predicted by using two artificial intelligence techniques such as
adaptive neuro-fuzzy inference system (ANFIS) and least square support vector
machine (LSSVM).
The methodology adopted in this study is quite general and can be extended to
design of hand tools, machinery, vehicles and furniture used in various work
environments. The numerical approach considered in this work may be extended to
dynamic analysis where vibrational effect can be analyzed in a moving vehicle.
Contouring of the seat can be considered to study its influence on pressure
distribution at ischial tuberosity. Kinematic model proposed in this study can be
extended to model the whole human body with more number of degrees of freedom.
Keywords: Ergonomic design; Subjective and objective design characteristics;
Comfort level; Artificial intelligence techniques; Bureau of Indian
Standard; Numerical model; Finite element model; Ischial tuberocity;
Multi-attribute decision making (MADM) approach; Kinematic model;
Neuro-fuzzy inference system (ANFIS); Least square support vector
machine (LSSVM).
iv
CONTENTS
Chapter Title Page No
No
Acknowledgement i
Abstract iii
Contents v
List of Tables vii
List of Figures ix
1 Background and Motivation

1.1 Introduction 1
1.2 Application of ergonomics 2
1.3 Need for research 3
1.4 Research objective 4
1.5Thesis outline 5
2 Literature Review
2.1 Introduction 10
2.2 Ergonomic consideration for improving user satisfaction 12
2.3 Bio-mechanical analysis in ergonomic design 15
2.4 Ergonomic consideration in product selection 17
2.5 Layout design for improving work environment 18
2.6 Conclusions 19
3 An Integrated Approach for Designing Office Furniture

with Ergonomic Consideration
3.1Introduction 21
3.2 Methodology 22
3.2.1 Data collection 22
3.2.2 Factor Analysis 25
3.2.3 Quality function deployment (QFD) 26
3.2.4 Adaptive Neuro Fuzzy Inference System (ANFIS) 26
3.2.5 Design of Experiment 30
3.3 Results and discussions 30
3.4 Conclusions 52
4 A Numerical Approach for Ergonomic Design

4.1 Introduction 54
4.2 Model descriptions 55
4.2.1 Material properties for seat 57
4.2.2 Material properties for human soft tissue 59
4.3.1 Model validation 62
4.4 Conclusions 74
5 A Novel Multi-Attribute Decision Making Approach for

Product Selection Conforming Ergonomic
Consideration
5.1 Introduction 75
5.2 Proposed methodology 76
5.2.1 Techniques for Order Preference by Similarity to 81
v
Identical Solution (TOPSIS)
5.2.2 VIseKriterijumska Optimizacija I Kompromisno 82
Resenjea (VIKOR)
5.2.3 Preference Ranking Organization Method for 84
Enrichment Evaluations (PROMETHEE)
5.4 Conclusions 99
6 Human Upper Arm Posture Prediction within

Isocomfort Work zone
6.1 Introduction 100
6.2 Model description 101
6.2.1 Denavit -Hartenberg Representation 102
6.3 Methods for determination of IK solutions 107
6.3.1 LSSVM architecture 107
6.3.2 ANFIS architecture 108
6.5 Conclusions 114
7 Executive Summary and Conclusions

7.1 Introduction 116
7.2 Summary of findings 117
7.3 Contribution of research work 118
7.4 Limitations of study 120
7.5 Scope of future research 120
Bibliography 121
Appendix
List of Publications a
vi
LIST OF TABLES
Table Caption Page
No. No.
2.1 Summary of publications referred 11
3.1 Customer requirements 23
3.2 Tangible design attributes 24
3.3 Intangible design attributes 24
3.4 Anthropometric standing dimension (mm) 24
3.5 Anthropometric sitting dimension (mm) 25
3.6 Factor Analysis for all items (0.702) 31
3.7 Design attributes for three different QFD models 32
3.8 Ranking of Design attributes for model-1(comfortness) 33
3.9 Ranking of Design attributes for model-2(balance) 33
3.10 Ranking of Design attributes for model-3(luxuriousness) 33
3.11 Levels of various designs attributes (parameters) 46
Experimental design using orthogonal array of l36 2 3 
3 10
3.12 47
3.13 Comparison of optimum parameter with Bureau of Indian 49
Standard (BIS)
3.14 Dimensions of prototype (excluding known control parameter) for 50
comparison
4.1 Material properties for Ogden hyper-foam (SAF 6060) for soft 58
cushion (E=200 kPa,  = 60 kg/m3)
4.2 Coefficients of Prony series parameters for cushion exhibiting 58
viscoelasticity
4.3 Material properties for Ogden hyper-foam for soft cushion (E=20 59
kPa,  = 40 kg/m3)
4.4 Coefficients of Prony series parameters for cushion exhibiting 59
viscoelasticity
4.5 Material parameters for hyperelastic material to define the soft 59
human tissue
5.1 Decision matrix 76
5.2 Linguistic rating for alternatives selection 87
5.3 Fuzzy numbers associated with alternatives 88
5.4 Aggregate fuzzy number of alternatives 89
5.5 Crisp ratings of alternatives 89
vii
5.6 Normalized crisp ratings 89
5.7 Objective weights of attribute 90
5.8 Pair wise comparison matrix. 90
5.9 Subjective weight design attributes 91
5.10 Integrated Weight calculation 92
5.11 Normalized decision matrix 92
5.12 Weighted normalized matrix for alternatives 93
5.13 C *i
Ranking index ( ) of alternatives 93
5.14 Ranking of alternatives considering integrated weight 93
5.15 The ranking and the compromise solutions 94
5.16 Preference function(Pi) resulting from the pair wise comparisons 95
of the six alternative with respect to criterion depth of cut.
5.17 Positive 

( A)  , negative  (A) and net flows (A) for the scenario 95
5.18 Ranking of alternatives considering integrated weight 95

5.19 Ranking order comparison 98
6.1 D-H parameters of human arm model 104
6.2 Joint angle range for comfort zone (Diffrient et al. 1985) 104
6.3 Prediction comparison between LSSVM and ANFIS 112
viii
LIST OF FIGURES
Table Caption Page

No. No.
2.1 Percentage of papers under each classification category 12
3.1 A typical architecture of ANFIS structure 29
3.2 QFD model for comfortness 34
3.3 QFD model for balance 35
3.4 QFD model for luxuriousness 36
3.5 ANFIS model structure:(a) 37,38
comfortness,(b)balance,(c)luxuriousness
3.6 Distribution of predicted and actual response training: (a) 39,40
comfortness,(b) balance,(c) luxuriousness
3.7 Distribution of predicted and actual response testing: (a) 40,41
comfortness, (b) balance, (c) luxuriousness
3.8 Surface plots: (a) comfortness (b) balance(c) 42,43
luxuriousness
3.9 Residual plots: (a) comfortness (b) balance(c) 43,44
luxuriousness
3.10 Normal probability curve of residual at 95% confidence 45,46
level:(a) comfortness (b) balance(c) luxuriousness
3.11 Effect of control design parameters on response 49
3.12(a) Orthographic Projection (All dimensions are in mm) 51
3.12(b) prototype of office chair with optimized design parameter 52
(All dimensions are in mm)
4.1 Model of seat cushion and buttock (soft tissue) 56
4.2 Finite element model of seat cushion and buttock (soft 57
tissue)
4.3 Variation of von Mises stress with increase in time for 60
different thickness of seat
4.4(a) Finite element model for rigid seat with magnification of 61
muscle region subjected to high stress
4.4(b) Stress distribution at ischial tuberocity for erect 62
immobilized continuous sitting on rigid seat of thickness of
80mm for 1800 sec (von Mises stress=37193Pa High
stress area=11.20 mm2)
4.4(c) Stress distribution at ischial tuberocity for erect 62
immobilized continuous sitting on rigid seat of thickness of
80 mm for 5400 sec (von Mises stress=37997 Pa High
4.4(d) Stress distribution at ischial tuberocity for erect 63
immobilized continuous sitting for rigid seat of thickness of
80 mm for 9000 sec (von Mises stress=38442 Pa High
4.5 Variation of von Mises stress over time for soft cushion of 64
(a) elastic modulus=200kPa and density= 60kg/m3 (b)
elastic modulus=20kPa and density=40kg/m3 for erect
immobilized continuous sitting
4.6(a) Finite element model for soft seat cushion (E=200kPa and 64
 =60kg/m3) with magnification of muscle region subjected
to high stress
4.6(b) Stress distribution at ischial tuberocity for erect 65
immobilized continuous sitting on soft cushion (E=200kPa
and  =60kg/m3) for 1800 sec (von Mises stress=22002Pa
ix
High stress area=1.24 mm2)
4.6(c ) Stress distribution at ischial tuberocity for erect 65
4.6(d) Stress distribution at ischial tuberocity for erect 66
4.7(a) Finite element model for soft seat cushion (E=20kPa and 67
 =40kg/m3) with magnification of muscle region
subjected to high stress
4.7(b) Stress distribution at Ischial tuberocity for erect 67
High stress area=0.17mm2)
4.7( c) Stress distribution at Ischial tuberocity for erect 68
4.7(d) Stress distribution at Ischial tuberocity for erect 68
and  =40kg/m3)for 9000 sec (von Mises stress=22914Pa
4.8 Postures (a) neutral position (b) left wise leaning (c) right- 69
wise leaning
4.9(a) Finite element model for immobilized continuous sitting 70
with tilting angle of 300 for soft cushion with elastic
modulus of 20kPa and density 40kg/m3
4.9(b) Stress distributions at ischial tuberocity on immobilized 70
continuous sitting with tilting angle of 300 for soft cushion
of elastic modulus of 20kPa and density 40kg/m3 (von
Mises stress=14632Pa Time=1800sec)
4.9(c) Stress distributions at ischial tuberocity on immobilized 70
4.9(d) Stress distributions at ischial tuberocity on immobilized 70
Mises stress=15975PaTime=5400sec)
4.9(e) Stress distributions at ischial tuberocity on immobilized 70
4.9(f) Stress distributions at ischial tuberocity on immobilized 71
4.9(g) Stress distributions at ischial tuberocity on immobilized 71
x
4.10(a) Finite element model for immobilized continuous sitting 71
with tilting angle of 300 for soft cushion with elastic
modulus of 20kPa and density 40kg/m3
4.10(b) Shear stress distribution at ischial tuberocity on 71
immobilized continuous sitting with tilting angle of 300 for
soft cushion of elastic modulus 20kPa and density
40kg/m3 (Shear stress=8438Pa Time=1800sec)
4.10(c ) Shear stress distribution at ischial tuberocity on 71
4.10(d) Shear stress distribution at ischial tuberocity on 72
40kg/m3 (Shear stress=8738Pa
Time=5400sec)
4.10(e) Shear stress distribution at ischial tuberocity on 72
40kg/m3 (Shear stress=8922PaTime=7200sec)
4.10(f) Shear stress distribution at ischial tuberocity on 72
4.10(g) Shear stress distribution at ischial tuberocity on 72
4.11 Effect of time duration in tilted (300)sitting on (a) von Mises 73
stress (b) shear stress for soft cushion(E= 20 kPa and
density =40 kg/m3)
5.1 Graphical representation of generic MADM model 77
5.2 Linguistic terms to fuzzy numbers conversion (5-point 78
scale)
5.3 Preference indices for a problem consisting of three 85
alternatives and four attributes
5.4 Office chair model for analysis 86
5.5 Integrated subjective objective weight with wo=0 and ws=1 96
5.6 Integrated subjective objective weight with wo=0.5 and 97
ws=0.5
5.7 Integrated subjective objective weight with wo=1 and ws=0 97
6.1 Human model arm representation(a) human 102
arm(b)Coordinate frame
6.2 The denavit Hartenberg notation (Mittal and Nagrath, 103
2007)
6.3 Workspace for 7-DOF redundant manipulator showing 109
human arm extremity
6.4 Surface plot for 1 ,  2 ,  3 ,  4 ,  5 ,  6 ,  7 111,112
6.5 Residual plot for 1 ,  2 ,  3 ,  4 ,  5 ,  6 ,  7 through LSSVM 113,114
and ANFIS
xi
CHAPTER 1
BACKGROUND AND
MOTIVATION
1.1 Introduction
It is imperative to focus on design and performance of work system for ensuring a
healthy and safe work environment that leads to improve productivity. Although
technological advancement leads to improve productivity, the risk factors emerging
from complex interaction of employees and elements of work system needs to be
sufficiently dealt in involving ergonomic concepts in order to enhance human
performance (Peterson,1997).Therefore, ergonomic design of a
product/equipment/machine becomes highly desirable because interaction between
user and product/machine leads to enhance user satisfaction, comfort and
performance of users minimizing health risks. Ergonomic typically solves the physical
problem associated with work environment by reducing mismatch between user
anthropometric and biomechanical parameters with physical dimension of work place,
equipment, furniture (Bridger, 1995; Jeong and Park, 1990). The physical problems
resulting due to anthropometric and biomechanical aspects are excess muscle loads,
posture change, working with awkward posture, exposure to constant static and
vibration force, repetition and duration of body movements. Such problems lead to
fatigue and musculoskeletal disorder and physical injuries in low back area, upper
extremities and blood circular system (Nunes, 2009). World Health Organization
(WHO) has recognized work related musculoskeletal disorder (WMSD) as a major
part of health problem that directly affects the behavior of employees and become an
influential factor for occupational accidents (Hilton and Whiteford, 2010; Sobeih et al.,
2009). It has been observed that repetitive work in the same posture can lead to
antagonistic of muscle tendon resulting in degradation of joint function (Bridger, 1995;
Marras and Schoenmarklin, 1993).
Consideration of ergonomic principles and methodologies not only involves
anthropometric and biomechanical aspects at design phase to improve the physical
capability of employees but also touches psychological needs of user to enhance
user satisfaction (Kuoppala et al., 2008; Morag, 2007; Moreau, 2003; Fredriksson et
al., 2001; Neumann et al., 2006; Vink et al., 2006; Kazmierczak et al., 2007; Erdinc
and Vayvay, 2008; Falck et al., 2010; Axelsson, 2000). User satisfaction associated
with psychological aspect and functional requirement of the product/machine
invariably improves the performance of the users and considered as an important
issue in designing a consumer/industrial product (Dandavate et al., 1996; Yang et al.,
1999; Park and Han, 2004). Improving functional requirements not necessarily satisfy
the user thoroughly especially in case of consumer products such as compact disc
(CD) players, mobiles, personal data assistant (PDA) etc. In such cases, subjective
performance of the product also matters to improve the user satisfaction. Usability of
1
a product comes through both functional requirement as well as subjective
performance of the product. Usability is defined as the degree of ease of use
(subjective performance) and effectiveness of use (objective performance or
functional requirement) within a specified users, tools and work environments
(Bennet, 1984; Shackel, 1984; Veryzer, 1995). Objective performance explains how
to interpret the product, how fast the users use/control the product and how well the
product functions. Subjective performance measures user perception of image and
impression regarding the product that explains the appearance of product as well as
the attitude or judgmental feelings about the product (Han et al., 2000).
1.2 Application of Ergonomic design
Ergonomics concern with human performance considering human physiology and
psychology for the improvement of work system consisting of person, jobs, tools,
equipment and workspace. Three general applications of ergonomics are observed in
practice as discussed below.
Ergonomics in industry
In manufacturing and service sectors, ergonomics reduce risk of work related
injuries and fatigue by designing the tools and equipment within user physical
capability so that the operator becomes flexible with work environment. Ergonomics
emphasize on design of the tools, equipment and work processes to improve work
productivity and efficiency providing safety (Lucas, 1984). Ergonomics usually
emphasize physical work load and productivity issues alongside safety and health
(Schmidtke, 1989; Luczak and Volpert, 1987).
Ergonomics in equipment design
Inadequate equipment operation leads to awkward/inaccurate posture and body
vibration giving rise to musculoskeletal disorder, unhealthy and unsafe work place
(Santos et al., 2011).The application of ergonomics can reduce the complex
operation on operator so that the task demand can be compatible with human
capabilities. Operator work schedule/habits should be considered during design of
the equipment to make the way of operation easy. The design of equipment is always
a compromise between the operator’s biological needs and physical requirements of
the equipment (Das and Grady, 1983; Das and Sengupta, 1996).
Ergonomics in product design
As the demand of a consumer product mostly depends upon the needs of end
user and producer cannot control the skill level of user, it is difficult for the part of
manufacturer to control over the end use of product. The ease of use of product is
associated with ergonomics (Weiman, 1982). Although product design based on
functionality, quality and cost as important factors, ergonomics emphasizes other
2
design elements such as comfort, safety (Vink et al., 2006), image/impression
(Jordan,1998), emotion and attractiveness (Nagamachi, 2002; Park and Han, 2004).
Ergonomic provides an opportunity to use the products easily and safely. For
example Volvo, Mercedes provides maximum comfort with their ergonomic design of
seat and layout. Also during the design of key board, mouse, electronics product and
furniture, ergonomic has a value to use the product easily.
1.3 Need for research
In spite of several applications of ergonomics design, a number of key barriers
still exist in many design processes. It becomes difficult to conceptualize the
relationship between subjective feelings and design characteristics. Since user
requirements are product and situation specific, it is difficult to predict the change in
subjective feelings. In addition to the issues related to user satisfaction, ergonomics
also concern with health related issues. Health injuries in work environment are
caused due to poor/awkward work posture, repetitive and continuous work for
prolonged time duration, incompatibility between user and tools/equipment. Health
risk factors when combined with poor machine structure, equipment, tools and
workspace create a physical and mental stress and fatigue on human body. As
muscle fatigue induces musculoskeletal disorder, it is important to quantify fatigue
and maximum limit of tolerable muscle loads (Burderf, 1992; Chaffin et al., 1999;
Armsstrong et al., 1990). In order to address to a large number of users satisfying
their necessity and preference, many products are available in the market. In such a
market environment, selection of a particular product becomes a difficult task. With
such ideas in mind, an effective method can be opted in the presence of multiple,
usually conflicting criteria to find out a suitable product with all important design
features. In addition to the issues related user comfort and satisfaction, ergonomics
also concern with design of workplace, layout and work facilities to balance the
interaction between human beings and tasks for providing safe and enabling
workplace environment (Schnauber,1986; Resnik and Zanotti,1997; Burri and
Helander,1991; Shikdar and Das,1995; Das and Sengupta,1996).
These problems offer new opportunities for ergonomic design as follows:

 Concept of user satisfaction should be emphasized from both subjective and
objective design requirements to evaluate a product’s usability (Nagamachi,
1995; Han et al., 2000).
 As it becomes a complex task to design a product/equipment/machine relating
design characteristics of product and user requirements, an appropriate
customer driven approach must be adopted.
3
 It is also necessary to develop a relationship between product characteristics
and user comfort level using an appropriate method due to non-linearity of the
relationship.
 It is necessary to investigate various physical exposures of muscles and
predict muscle fatigue in the work environment using as simple and efficient
tool.
 The physical discomfort is usually associated with biomechanical aspects
(Anderson et al., 1979). Hence, biomechanical analysis using analytical or
numerical models can help to quantify the damage on soft tissues of the
users, fatigue on human limbs and musculoskeletal disorders occurring
through interaction of product/machine and human body (Tewari, and Prasad,
2000; Thakurta et al., 1995).
 Although a product can be designed considering subjective and objective
design requirements and health related risk factors, it is often difficult to
choose a particular product in a market place with selected functions
satisfying large number of users. Therefore, a suitable selection approach in a
well-structured manner must be developed to provide a feasible alternative in
the presence of multiple and conflicting criteria.
 Always poor working posture, unnatural postures and irregular motions have
been considered as major cause of musculoskeletal disorders (Haslegrave,
1994). Therefore, it is important to design a comfort work zone for users to
assume good working posture for task performance minimizing stress and
discomfort. Improved comfort work zone enables the operators to use the
hands correctly and safely reducing unhealthy and lengthy reaches.
1.4 Research objectives
Based on discussion in previous sections, this section of thesis addresses the
issues and problems related to system design. Present work focuses a framework
providing the guidelines and principles to the designer and decision makers to
improve the quality, usability of product (office chair) in order to increase the comfort
level, reduce the stress and injury due to prolong sitting in office work environment.
The objective of the research is to develop the models by extracting the guidelines in
order to reduce the risk of musculoskeletal disorder in office environment and
increase customer satisfaction and hence productivity.
Based on this guiding principle, the objective of present work are as follows:
4
 To develop an integrated approach to handle objective and subjective user
requirements while designing office furniture with a view to enhance customer
satisfaction.
 To develop stress-time relationship on human muscle tissue through
numerical analysis and ascertain the cause of pressure ulcer so that cell
death can be minimized.
 To propose a multi-attribute decision making approach for selecting a suitable
ergonomically designed product under uncertain decision making
environment.
 To predict the comfort work zone considering comfort posture within which
materials and controls existing in the work place are easily accessible.
1.5 Thesis outline
To meet the above the objectives, the thesis can be organized into seven chapters
including introduction. A brief outline of each chapter is given as follows:
Chapter 2: Literature review
The chapter deals with review of related literatures that provide background
information on the issues and problems to be considered in the thesis and hence
focus the relevance of the present study. It identifies the problems associated with
work system design with relevance to technical and psychological aspects, health
and safety issues, comfort and productivity. The chapter proposed different
approaches to identify various aspects in work system design. Present research
summarizes on various system design issues considering ergonomics guidelines
Chapter3: An integrated approach for designing office furniture with ergonomic
considerations
Since it is difficult to manage subjective requirements in the design process, this
chapter proposes an integrated approach to deal with subjective and objective design
criteria for a product with ergonomic consideration. The procedure is demonstrated
with the help of design of an office chair. A questionnaire survey was conducted
including information on user requirements regarding office chair, anthropometric
sitting and standing dimensions and design parameters of office chair. The survey
was conducted through different modes over a period of four months. The
respondents are advised to provide ratings in a five point Likert type scale (1 for
strongly agree and 5 for strongly disagree) on forty different items. The user
requirements and design criteria are suitable to the users having average
anthropometric dimension. A total of one hundred fifty responses are collected. Sixty
percent of the respondents belong to officers and staffs of different banks and twenty
5
percent respondents belong to technical institutes. Rest data are collected from
different government/private offices. After screening, 100 data with useful responses
are considered for further analysis. Data reduction technique like factor analysis has
been applied to survey data in order to eliminate redundancy. Thirteen different
models of office chairs from standard manufacturers are displayed to the respondents
to finalize the tangible and intangible design parameters of the office chair. Total of
forty different adjectives indicating user requirements were identified. Twenty two
user requirements were loaded on three factors having factor loading score more
than 0.7.The reduced user requirements are translated into design characteristics
using quality function deployment (QFD). The nonlinear relationship between design
characteristics and user satisfaction is developed through adaptive neuro-fuzzy
inference system (ANFIS). Finally, a large number of design scenarios are generated
using design of experiment (DOE) approach and the best design parameters are
chosen that maximizes user satisfaction. A prototype model is developed by using
optimal design parameters. The design parameters are compared with
measurements by Bureau of Indian Standards (BIS). The comparison indicates that
some of the design parameters of the proposed model are out of range suggesting
that the variation is due to localization of sample data. The standard needs to be
reviewed regularly to enhance comfort level of the users.
Chapter 4: Study on human-product interaction by means of a stress analysis to
minimize risk of injuries
Once the prototype with proper design parameters satisfying user satisfaction has
been developed, the numerical model is developed to analyze the effect on human
tissues due to prolonged use. The model includes a soft human tissue, ischial
tuberosity (a bony part) and the seat cushion/rigid seat. Soft tissue (buttock)-seat
model is assumed to be two dimensional axisymmetric finite element model with an
upright posture. The nonlinear soft tissue-seat model is developed using ANSYS
10.0. A volunteer weighing fifty five kilogram is contacted to obtain image of the soft
tissue (buttock). The seat cushion has been modeled by a rectangular flat surface
having thickness of 80 mm and area of 450×450 mm2. Two different models
comprising of human soft tissue-rigid seat (model I) and human soft tissue-soft
polyurethane foam cushion (model II) are considered. The simulation is conducted
over a period of three hours to study the effect of time of loading on soft tissue. The
stress distribution throughout meshed model is studied by varying the material
properties of seat cushion, angle of loading (sitting posture) and cushion thickness. It
is observed that the maximum stress affected area is less in case of soft cushion
6
having elastic modulus of E=20kPa and density  =40 kg/m3 than the cushion having
elastic modulus of E=200kPa and density  =60kg/m3. The model predicts a

maximum stress of 20324Pa at ischial tuberosity after a continuous sitting duration of
half an hour on a soft cushion having elastic modulus of E=20kPa and density  =40
kg/m3 and comparable with experimental value (nearly 19500Pa) (Verver et al.,
2004). The morphological changes at ischial tuberosity has been noticed as a result
of von Mises stress in soft muscle tissues for three different cushion types with
continuous sitting for 1800sec. It has been seen that the size of damage area
decreases from rigid seat to soft cushion. It is minimum for the cushion having elastic
modulus 20kPa and density 40 kg/m3. The trend of increasing stress with increase in
time of sitting at ischial tuberosity is nearly similar for all seat types. The analysis also
been done by changing the posture from 00 to 300 to investigate the effect of loading
direction and time of loading on change in stress and damage area at ischial
tuberocity.
Chapter 5: A novel multi-attribute decision making approach for product selection
conforming ergonomic considerations
In the market place, a large number products are available with various design
features (design attributes) with ergonomic considerations. However, it is difficult to
choose a particular one suiting to user’s needs. Multi attribute decision making
(MADM) approach provides a structured approach in selecting a feasible alternative
in the presence of multiple and conflicting criteria. Generally, MADM approach carries
objective weight for design attributes to find a feasible alternative. But the weights
need not be necessarily expressed objectively. In many situations, the weights are
expresses objectively and subjectively. In this chapter, a novel decision making
approach has been explored to consider both subjective and objective weights of
design attributes so that the decision maker facilitates with the objective information
regarding the product as well as the uncertainty of human judgement. The approach
is explained with an example having six alternatives based on ten attributes of
different design features (attributes).
The structure of the present work is as follows. Section 1 concerns with finding
the design characteristics (attribute) with respect to alternatives (office chair). Section
2 considers both objective subjective weights of various attributes. On the basis of
statistical variance method, the objective weights of the attributes are computed.
Analytic Hierarchy Process (AHP) is used to calculate subjective weights of
attributes. The integrated weights of attributes are obtained considering the different
weightings proportion of the objective and subjective weights. Four decision makers
7
are involved to assign the fuzzy rating values to alternatives under each attribute
using a five-point fuzzy scale with triangular membership functions. In order to assess
the rating of alternative under each attribute, the fuzzy numbers are aggregated. The
aggregated fuzzy rating values are converted into crisp values to simplify the
calculations using left and right score (Chen, 1985). The aggregate crisp values are
then normalized. The normalization is based upon beneficial and non beneficial
attribute. For beneficial attributes higher values are desired (maximization) whereas
lower values (minimization) are preferred for non-beneficial attributes. In Section 3,
three different MADM methodologies such as TOPSIS (Techniques for Order
Preference by Similarity to Ideal Solution), VIKOR (VIseKriterijumska Optimizacija I
Kompromisno Resenjea) and PROMETHEE (Preference Ranking Organization
Method for Enrichment Evaluations) are utilized to find out the best possible way to
choose the suitable office chair. In order to check the stability of the ranking, a
sensitivity analysis has been carried out considering different proportion of attribute
weights (subjective and objective).
Chapter 6: Kinematic analysis of human upper extremity based on comfort joint
posture reducing unhealthy and awkward posture.
In this chapter, a kinematic model of human upper extremity is analyzed allowing
the movements of all axes of kinematic chain within a comfort zone so that fatigue on
human limb can be reduced. The diagnosis of posture analysis of the upper
extremities within the comfort operating zone allows the operator to have a comfort
work range within which possible posture can be accepted. With the help of comfort
joint angle range from literature (Diffrient et al., 1985) as comfort posture and different
segment measurements of upper arm (link length) (Kaur et al., 2011;Singh et al.,
2013), forward kinematic equations are developed to achieve the hand reach position
and consequently three dimensional comfort work zone. The parameters used in the
link segment are extracted from the model of Murray (Murray, and Johnson, 2004).
Human body carries a number of links with offset joints and adapts some specific
postures to formulate specific task. Therefore, it is important to get a right posture
within comfort work zone. Due to the presence of non-linearity, complexity and
singularity issues in solving inverse kinematic problem, two artificial intelligence
techniques such as adaptive neuro-fuzzy inference system (ANFIS) and least square
support vector machine (LSSVM) are used to predict upper arm posture (comfort joint
angles). It has been seen that LSSVM shows a better performance with less error
(0.1173428) in comparison to ANFIS (0.506631) for predicting the posture.
8
Chapter 7: Executive summary and conclusion
This chapter presents the summery of the results, suggestions and scope for
future work in the direction of ergonomic design. With specific contribution and
limitations, the chapter concludes the work enclosed in the thesis.
9
CHAPTER 2
LITERATURE REVIEW
2.1 Introduction
Demand for improved performance of product and comfort for the operator/user
have led to an increasing emphasis on ergonomic design. Development of a
comfortable product bridges the gap between the subjective feeling of comfort and
the prediction of comfort level of new designs through practical use. Various
customer-driven approaches allow the needs of the customer to be communicated
through the various stages of product planning, design, engineering and
manufacturing into a final product to improve the performance of product/machine
and user satisfaction. Sometimes, simulation of computer models to test the design
process for highest degree of comfort allows manufacturers to speed up the design
process by reducing cost and minimizing health hazards. Due to advanced
technology, convenient functions are added to a product. However, providing more
functions usually results in a more complex user interface which sometimes neither
offer user satisfaction as a part of usability nor provide physical comfort. Customer
demand for product with improved performance is as important as the demand for
products with improved comfort level. Therefore, the manufacturers consider user
satisfaction, comfort and biomechanical issues as the most important factors that
distinguish each product/tool/equipment from each other product/tool/equipment.
Ergonomics not only concern with health related issues (Kuoppala et al., 2008) and
psychological aspects (Fredriksson et al., 2001) but also associated with system
performance aspects like productivity (Kazmierczak et al., 2007) and quality (Falck et
al., 2010) by integrating various features in the product t design stage.
In this direction, the present chapter highlights various concepts and approaches
through a broad-based literature survey. Current literature survey deals with ninety
articles published after 1992 with attention paid to last twenty two years. Sixteen
articles referred here are published before 2000 and rest published after 2000. The
majority of the citations are found from peer reviewed journal publications (90%). The
other citations are taken from conference articles and book chapters. Two journals
namely International Journal of Industrial Ergonomics and Applied Ergonomics
together account for 35% of total citations in journals. Table 2.1 provides the source
and number of citations from each source. The literature review provides enough
information regarding existing problem to identify gap in the existing work and
provides advancement in solving the problem and minimizing the gap so that the
relevance of the present work can be emphasized.
10
Table 2.1 Summary of publications referred
Name of Journals Citation
International Journal of Industrial Ergonomics 24
Applied Ergonomics 15
Journal of Occupational and Environmental Medicine 1
The Ergonomics Open Journal 1
International Journal of Industrial Engineering Computations 1
Expert System with Applications 3
Applied Soft Computing 1
Mathematical and Computer Modeling 1
Journal of Human Ecology 1
International Journal of Services and Operation Management 1
Journal of Convergence Information Technology 1
Computers and Industrial Engineering 1
Materials and Design 1
Journal of Biomechanics 4
Journal of Engineering 1
Computer Methods in Biomechanics and Biomedical Engineering 1
International Journal of Solids and Structures 1
Cellular Polymers 1
Journal of Rehabilitation Research and Development 3
Journal of Electromyography and Kinesiology 1
International Journals of Simulation Modeling 1
Journal of Biomechanical Engineering 3
Archives of Physical Medicine and Rehabilitation 1
Computers in Industry 2
Journal of Mechanical Design 1
Journal of Operation Management 1
Human Factors and Ergonomics in Manufacturing and Service 3
Industries
Ergonomics 2
Journal of the Chinese Institute of Industrial Engineers 1
Computer Standards and Interfaces 1
Scandinavian Journal of Work, Environment and Health 1
International Review of Social Sciences and Humanities 1
Journal of Electronic Commerce Research 1
The Journal of Visualization and Computer Animation 1
Conferences 6
Books 1
Total 91
The referred literature is broadly classified into four categories such as (1)
Ergonomic consideration improving user satisfaction, (2) Biomechanical analysis in
ergonomic design, (3) Ergonomic consideration in product selection and (4) Layout
design for improving work environment as illustrated. Each classification refers work
system design issue with ergonomic consideration and associated with specific
problem. Figure 2.1illustrates percentage of papers surveyed under each
classification category. The next sections provide a brief discussion on these issues.
11
Literature Review
Ergonomic consideration
improving user satisfaction
15%
40% Biomechanical analysis in
14% ergonomic design
Ergonomic consideration in
31%
product selection
Layout design for

improving work
environment
Figure 2.1 Percentage of papers under each classification category
2.2 Ergonomic consideration for improving user satisfaction

User satisfaction generally depends on perceived image/impression, functionality
and physical comfort during user-product interaction. To relate user satisfaction with
product’s design features, various research studies focus on embedding affective
human factors into product design. Usually, mismatch between product and user
anthropometric data can lead to an uncomfortable and awkward body posture making
adverse impact on working condition. Keeping in view with user comfort, attempts
have been made to embed customers’ views in product design through interviews
and well-structured questionnaire survey in order to design the product to
accommodate anthropometric variability. Reitenbach et al. (2009) have proposed an
office chair design to support the postures of Chinese female office workers in a
comfortable way by means of interviews, questionnaire survey and focus group
discussions. Focusing on the comfort level of user and usability of office chair, the
study conducts a comparative analysis on Hong Kong office workers and Chinese
factory workers. Mokdad and Ansari (2009) have conducted a survey on Bahraini
school students and suggested variability in design of furniture for them by comparing
the anthropometric dimensions of boys and girls (6-12 years).Goonetilleke and
Feizhou (2001) have proposed integration of subjective and objective measurement
to evaluate the useful seat depth for a target population. Objective evaluation
revealed that a seat depth of 31-33 cm is adequate for South China region population
whereas seat depth of 38-43cm is adequate for US population. Considering
12
anthropometric data as a major parameters for product design, many countries have
been making great efforts in establishing an anthropometric database for different
population groups such as civilians, military personnel, students and workers (Bolstad
et al., 2001; Wang et al., 2002).Thariq et al. (2010)have evaluated design parameters
(dimensions) for fixed type side mounted desktop chair to improve comfort level by
developing relation between the chair dimensions and the desk dimensions with
anthropometric dimension of university students of Sri Lanka. Musa (2011) has made
a comparison between students of different age groups by their anthropometric
dimensions and concluded that the design of furniture for 12-year students does not
match with that of 17-year students. The analysis emphasized in selecting different
design criteria for different age groups so that there is less chance of mismatch
between school furniture and students’ anthropometric dimension. Jindo and
Hirasago (1997) have described the style and design specification of passenger car
interiors by subjective evaluation and proposed empirical relationship using multiple
regression method. Gauvali and Boudolos (2006) have utilized theoretical and
practical ergonomic principles to provide a relation between furniture dimensions and
anthropometric measures for choosing unique furniture dimension for unique
anthropometric measurement. Castellucci (2010) has made a comparison of various
furniture sizes of different schools based on anthropometric data of Chilean students
in Valparaiso region. It is found that mismatch exists between furniture dimension
(seat height and seat to desk height) and students’ anthropometry. Jung (2005) has
designed a new structure with minimum controls, cost and maximum flexibility of a
prototype of adjustable table and chair. The prototype design is validated considering
students’ physical dimensions through subjective trials and the dimensions provided
by International Standards data. Lin and Kang (2000) present an anthropometric
database for designing high school and primary school desks and chairs keeping in
view with subjective comfort level and anthropometric data. Ray et al. (1995) have
presented a statistical analysis of anthropometric data of Indian school students to
facilitate the furniture and toys that reduce biomechanical and visual problems.
Several studies focus on various customer driven approaches into product
development process in order to translate design features into user needs aimed at
maximizing user satisfaction. Han et al. (2000) have emphasized the importance of
objective performance and subjective impression during design phase to explain the
usability of product and identified the relationship between design variables of
electronics product and usability using multiple linear regression techniques.
Menendez et al. (2011) have studied ergonomic aspects of a highly adjustable office
chair and its impact on employees through statistical analysis. Shimizu and Jindo
13
(1995) have proposed a method for analyzing sensitivity evaluation using a fuzzy
regression method which takes into account the non-linearity of human sensation in
designing vehicle interiors. Khanam et al. (2006) have explored the type of furniture
used by the graduate students in the classroom environment keeping in view with
relationship between physical design, physical structure and biomechanics of human
body. Chen and Ko (2008) have proposed a fuzzy quality function deployment (QFD)
approach for designing new product that maximize the customer satisfaction. The
proposed model is illustrated with a case study from semiconductor packing industry.
Jindo et al. (1995) have focused on design support system for design of office chairs
and conducted a subjective evaluation using semantic differential method in order to
find out a relation between design elements and user perception. Nagamachi (1995)
in Kansei engineering approach has proposed an elegant methodology for translating
customer requirements expressed in subjective manner into objective design
attributes using statistical method. Solomani and Zhong sheng (2006) have prioritized
the design characteristics using QFD.
Vergara and Page (2000) have studied the relation between lumbar and pelvic
posture with the backrest of a chair through extensive experimentation. Park et al.
(2000) have proposed new design of a chair to reduce muscle fatigue and discomfort
as compared to conventional computer chair. Vos et al. (2006) have investigated
experimentally the impact of postural and chair design on seat pan interface
pressure. It is concluded that the chair design has greater impact on seat pan
interface pressure than postural change. Ellegast et al. (2012) have evaluated the
effect of task performed during sitting on a chair on human muscle. Groenesteijn et
al. (2009) have investigated the influence of chair parameters on comfort and seat
interface pressure when prolonged work is being done using office chair. Lili et al.
(2010) have provided guidelines on design of office chair using ergonomics to
improve the comfort level of the users.
In order to quantify vague nature of expression on comfort by the users, several
studies have used fuzzy models to build the relationship between design
characteristics and user requirements. Dursun Kaya et al. (2003) and Kwong et al.
(2009) have adopted adaptive neuro-fuzzy inference system (ANFIS) to develop
relationship between user satisfaction and design parameters of school chair and
desk. The design parameters are established through six anthropometric dimensions
of boys and girls. Wang (2011) has developed an approach based on rough set
theory, ANFIS and Kansei engineering to convert customers’ preferences into
product form elements. The investigation has considered some adjectives to describe
the consumers’ psychological feeling on a product. Park and Han (2004) have used a
14
fuzzy rule-based approach to build the relationship between design variables of
products and user satisfaction. Three different chair dimensions such as
luxuriousness, balance and attractiveness with a number of continuous and
discontinuous variables are considered to build the model. To verify the model
performance, traditional regression model is compared with proposed fuzzy model.
Jiang et al. (2012) have attempted to model the relationship between user
satisfaction and design attributes of products using swarm optimization in conjunction
with ANFIS. In order to find the effectiveness of the proposed approach, modeling
results are compared with fuzzy regression and genetic algorithm based ANFIS
approach. Chan et al. (2011) have used genetic programming (GP) based QFD to
develop relationship between engineering characteristics and customer requirement
for a digital camera. The method is compared with linear regression and fuzzy
regression approach.
2.3 Bio-mechanical analysis in ergonomic design

Human psychological conditions (comfort or discomfort) are always associated
with body’s biomechanical and physiological perspective. Feelings of discomfort and
uneasiness are associated with tiredness and pain whereas comfort is associated
with relaxation (Helander and Zhang, 1997).Biomechanical problems caused by the
product due to lack of capabilities to perform the tasks need to be identified to
prescribe preventive measures in injury-prone situations. As muscle fatigue leads to
cumulative trauma disorder (CTDS), it is important to quantify fatigue and identify
maximum muscle load that a human body can tolerate (Chaffin et al., 1999). Buckle
and Devereux (2002) have defined musculoskeletal disorder (MSD) in terms of
injuries at muscles, joints, ligaments and cartilage caused due to repetitive task.
Factors such as awkward posture, prolonged contact stress, forceful exertions,
vibration and environmental conditions cause MSD. With time, it converts from mild
symptoms to severe chronic conditions. Visual discomfort and musculoskeletal
discomfort in neck and shoulders are most common occupational health concerns for
people who work with computers continuously for a prolonged period of time
(Berggvist and Knave, 1994; Hunting et al., 1981).
Ma et al. (2008) have proposed a seat-buttock model to find out the contact
pressure at seat-buttock interface and stress at bony prominence. Gru-jicis et al.
(2009) have examined stress distribution over the seated-human/seat interface
through a realistic model of car seat. Siefert et al. (2008) have proposed a human
model interacted with car seat to estimate seat pressure distribution using finite
15
element analysis. Fok and Chou (2009) have proposed a human finger model in
order to predict the internal loading pattern at tendons and joint surfaces during
dynamic motion under different flexion-extension joint angles. Tang et al. (2010) have
considered a two-dimensional human buttock-thigh model to investigate the effect of
varying vertical vibration frequencies on seat-interface contact pressure during sitting
on three different seat cushions using a finite element model.
Considering pressure distribution as an objective measure of discomfort, various
researchers have proposed relations between chair seat pressure distribution and
comfort level of the user. Verver et al. (2004) have proposed a finite element (FE)
model of the human buttocks to predict the pressure distribution between human and
seating surface with detailed and realistic geometric description. A parametric study
indicates that a pressure distribution at human-seat interface strongly depends on
variations in human flesh and seat cushion properties. Wang and Lakes (2002) have
analytically investigated the contact problems between two homogeneous and
isotropic soft bodies to simulate the contact of human buttocks and seat cushions
allowing Poisson’s ratio of seat cushion to be negative. Analysis by both the Hertz
model and a finite thickness 3D elasticity model shows that cushions with negative
Poisson’s ratio can reduce the contact pressure and prevent pressure-induced
discomfort and pressure sores/ulcers in sick people. Lowe and Lakes (2000) have
reached a similar conclusion using a FE model. Moes and Horvath (2002) have
proposed a FE approach for shape optimization of seats considering interactive force
between seat and body. Hobson (1992) has studied the effect of seated posture and
body orientation on pressure distribution and shear force acting at body seat interface
within and between two study groups made up of subjects with spinal cord injuries
and nondisabled subjects. Silver-Thorn and Childress (2003) have investigated the
effect of parameter variations on the prosthetic interface stresses for persons with
trans-tibial amputation using FE approach.
As most of the manual work is still done with hand tools, badly designed tools can
induce upper extremity musculoskeletal disorder like hand-arm vibration syndrome
and carpel tunnel syndrome (Punnett and wegman, 2004).Harih and Dolsak (2013)
have developed hand tool handles which can avoid deformation in soft tissue due to
higher contact area and anatomical shape of handles. Vignais and Marin (2014) have
proposed a biomechanical analysis of upper arm extremity during cylinder grasping
based on inverse kinematics. In an effort to reduce the incidence of decubitus ulcers
among wheelchair users, Todd and Thacker (1994) have emphasized on cushion
design to minimize the pressure at the buttock-cushion interface using finite element
analysis. Linder-Ganz et al. (2006) have used FE approach for estimating tissue
16
deformation over critical time durations causing pressure injuries using muscle tissue
of albino (Sprague–Dawley) rats exposed to pressures. Ceelen et al. (2008) have
conducted experiments using magnetic resonance (MR) and T2-weighted MR
imaging to measure the tissue deformation and damage. A finite element model is
proposed to calculate the strain in damage experiment. A correlation analysis
revealed a linear correlation between experimental and numerical strains. Gefan et
al. (2008) have indicated through specialized experiments on planar tissues that
there is 95% likelihood that cells could tolerate engineering strains below 65% for one
hour whereas the cells could endure strains below 40% over a 285 min trial period.
The decrease in endurance of the cells to compressive strains occurs between one-
three hour post-loadings. In another paper, Gefan et al. (2005) have proved that
stiffening occurs in-vivo in muscular tissue which undergoes widespread cell death
produced by applied bone compression. The local cell-death related stiffening affects
the distribution of mechanical stresses and deformations in adjacent (not yet
damaged) muscular tissue promoting deep pressure sore. Linder-Ganz and Gefen
(2009) have pointed out that the efficacy of wheelchair cushions should be evaluated
not only based on their performance in redistributing interface pressures but also
according to their effects on stress concentrations in deep tissues, particularly
muscles to minimize deep tissue injury. Ragan et al. (2002) have analyzed the effect
of cushion thickness on subcutaneous pressure during seating using finite element
modeling approach. Fathallah (2010) have focused on MSD in agricultural workers
considering psychosocial and socio-cultural aspect of the work environment.
2.4 Ergonomic consideration in product selection

In order to compete in the market-place, functions are being added to the product
based on assumption that more functions would enhance the product performance.
However, addition of more functions to a product usually results in complex user-
product interface making it difficult to use. Hence, it is vital to prioritize the important
features possessing the ability to fulfill use requirements during product design stage
(Besharati et al., 2006). Han et al. (2004), Chuang et al. (2001) and Han (2003) have
investigated the relationship between user satisfaction and design features of
different mobile phones. Park and Kim (1998) have used modified House of Quality
(HOQ) for selecting a set of design requirements of indoor building in order to
improve the quality of air. Park et al. (2011) have proposed a combination of three
approaches such as general usability principle, user interface component and
guideline properties to choose a suitable mobile. Lin et al. (2008) have proposed a
17
framework of analytic hierarchy process (AHP) and technique for order preference by
similarity to ideal solution (TOPSIS) to choose a suitable personal data assistant. Lin
et al. (2007) have presented a grey relational analysis approach for determining the
best combination of features in a mobile phone. Recently, Hua et al. (2014) have
proposed a hybrid multi-criteria decision making (MCDM) model like VIKOR
(Visekriterijums kokompromisno rangiranje) to select a smart phone. Isıklar and
Buyukozkan (2007) have proposed a MCDM approach to evaluate mobile phone in
accordance to user preferences. Liu et al. (2012) have focused on customer utility
generation, an optimum design selection approach based on fuzzy set decision-
making to identify design attributes from customer preferences using an analytical
hierarchy process. A multi-attribute analysis is developed to investigate the
preference of each attribute from the expert’s group decision. Conjoint analysis is
used in the product customization to find the effectiveness of model. Mokhlis and
Yaakop (2012) have focused on importance of different choice criteria (innovative
features, image, price, personal recommendation, durability, portable aspects, media
influence and post-sales service) in mobile phone selection among Malaysian
consumers. Guan and Lin (2001) have proposed a neural network approach to select
mobile phones.
2.5 Layout design for improving work environment

Layout design in a workplace enables the operator to use their hands and legs
safely and properly preventing unhealthy and awkward movements of body parts.
There may be possibility to apply maximum effort or may require extended reach to
achieve a specific task. Layout design needs understanding of human posture as well
as movements during work activities with maximum capabilities to obtain a safe,
healthy and comfortable work environment. Cimino et al. (2009) have proposed a
methodology based on 3-D simulation to evaluate the impact of workstation
parameters on multiple performance measures (force level to lift the objects, stress
level related to working posture, energy expenditure and process time). Hu et al.
(2010) have presented an experimental analysis on drilling task to estimate three
objectives performances such as maximum elbow angle extension, maximum muscle
force capacity reduction and task completion time and two subjective feelings like
discomfort in body parts and perceived exertion.
Kumar et al. (2009) have explored a tractor control layout (steering, foot clutch,
foot break and foot accelerators in the workspace envelop) for Indian people
considering anthropometric dimensions. Margarities and Marmaras (2006) have
18
suggested ergonomic requirements for individual work stations in an office
environment considering office equipment, environmental conditions, work
performance and usability related issues.
With the help of a CAD model, Rajan et al. (1999) have developed an integrated
virtual-reality based environment to analyze the assembly of product and jig design in
order to meet the required tolerance in aircraft industry. Dewangan et al. (2010) have
developed a statistical approach considering factor analysis to design agricultural
hand tools and equipment for workers in the hilly region of North East India.
Although layout design facilitates the task of the operator by positioning the
equipments/tools around, human posture prediction is also one of the most important
issues to determine a healthy work environment. Lindegard et al. (2005) have
developed a relation between VDU-user’s comfort rating and observed working
posture Comfort rating can be obtained through a questionnaire and the working
posture can be observed by an ergonomist. In order to simulate human posture (set
of joint angles) from a defined work zone (position of human limbs), researchers have
attempted to solve inverse kinematic solution through algebraic (Zhao and Badler,
1994), iterative (Jung et al., 1995) and analytical method (Hingtgen et al., 2004).
Wang and Verriest (1998 a) have proposed a geometric method for four degree of
freedom arm model to predict reach posture. They have investigated motion analysis
based on analytical inverse kinematic solution which is task oriented. Wang (1999)
have determined motion prediction for two activities such as serving water from a jar
and picking up a bottle.
2.6 Conclusions
Critical analysis of articles published in last few years in the broad spectrum of
ergonomics reveal that systematic framework is needed in design confirming to ease
human in human and products/machines/equipment/components interaction. The
interactions are sometimes expressed in both subjective and objective manner. The
objective characteristics relates to usability aspects whereas subjective
characteristics aims at enhancing satisfaction of the users. A broad framework is vital
to improve the comfort level of the users and subsequently productivity of the
organization. In addition to design framework, simulation of user-product model using
analytical and numerical analysis helps the designer to predict comfort level in terms
of stresses and fatigue being developed on human soft tissues/muscles during the
interaction process. Although empirical models and artificial intelligence techniques
are capable of developing relationship between important design variables and user
19
comfort level, simulation models referring to biomechanical analysis can be employed
to understand the physical underpinnings of interactions. The design analysis leads
to develop a user-friendly product. However, selection of best product out of a large
number of alternatives available in a specific situation needs managerial decision
making approach. Literature suggests that a large number of articles focus in
this direction. However, human judgement in a uncertain situation needs
careful application of various approaches available in the literature. During the use
of a product, estimation of comfort work zone is an important issue as evident
from the literature.
20
CHAPTER 3
AN INTEGRATED APPROACH
FOR DESIGNING OFFICE
FURNITURE WITH
ERGONOMIC
CONSIDERATION
3.1 Introduction
Ergonomically designed industrial or office work environment consider both

psychological and physical needs of employees during design phase for increasing
the job satisfaction and prevent the injuries in workplace (Braun et al., 1996). Since
the employees spend most of the time at the workplace, ergonomically designed
furniture plays an important role in decreasing fatigue and injury level even if an
employee continues prolonged work. This is vital for enhancing employees’ efficiency
and productivity in workspace. Nowadays, employees in office environment not only
engaged in studying and signing files but also work in computers, present business
plans, and discuss with colleagues. As a result, the employees sustain muscular
disorder and spine stiffness problems due to constant pressure at back, shoulder and
neck muscles (Kingma and Dieen, 2009). Groenesteijn et al. (2012) have evaluated
effect of office tasks on the posture and movements in different office environments.
In order to prevent muscular disorder, movement of muscle and spine should be
increased during sitting posture (Andreas et al., 2007). Since human-product
interaction focuses on subjective satisfaction as well as on objective performance,
studies have been carried out in the past to establish the relationship between user
sensitivities and design elements of office chair (Jindo et al., 1995). In the process of
enhancing customer satisfaction, more functions are added to the product to ensure
satisfaction in terms of convenience and ease of use, the structure of the product
becomes complex one. Therefore, human-product interaction is viewed as vital
element for product design. Usability of the product is concerned with the process of
use (i.e., how the user complete the tasks using system functions) measured in terms
of efficiency, effectiveness and satisfaction (Han et al., 2000). Kansei engineering
approach proposes an elegant methodology for translating customer requirements of
a product expressed in subjective manner into objective design attributes using
statistical method (Nagamachi, 1995; Horiguchi and Suetomi, 1995). In many
situations, multiple regression techniques are used to establish the relationship
between usability and design elements (Han et al., 2000). Optimization techniques
have also been used to find out best values of design parameters that maximize
customer satisfaction (Hong et al., 2002). In many cases, a functional model is
needed to describe the relationship between subjective customer requirement and
design elements.
Since it is difficult to manage subjective requirements in the design process,
present work proposes an integrated approach to deal with subjective and objective
design criteria for product development with ergonomic consideration. The procedure
21
is demonstrated with the help of design of office chair. Customers’ expectation from
the product is extracted through a questionnaire survey. Data reduction technique like
factor analysis has been applied to survey data to eliminate redundancy. The
reduced customer requirements are translated into design characteristics using
quality function deployment (QFD). The relationship between design characteristics
and customer satisfaction is developed through adaptive neuro-fuzzy inference
system (ANFIS). Finally, a large number of design scenarios are generated using
design of experiment (DOE) approach and best design parameters are chosen that
maximizes customer satisfaction. A prototype model is developed using the optimal
parameters and compared with Bureau of Indian Standards (BIS). The verification
result suggests that the proposed model parameters are within the prescribed ranges
of BIS.
3.2 Methodology
The complete methodology for designing an office chair with ergonomic
considerations can be explained with the help of following six steps. A well-structured
questionnaire is prepared to extract data on customer attributes, design
characteristics and anthropometric dimensions of users through cross-sectional
survey study. Factor analysis is carried out on customer attributes to reduce the
number of variables removing redundancy. QFD is employed to transform the
customer requirements into important design attributes. Since it is difficult to establish
relationship between design requirements and customer satisfaction due to
involvement of subjectivity, a black-box type predictive approach such as ANFIS is
used to map design requirements with customer satisfaction. Once ANFIS model is
well trained, good number of scenarios are generated using DOE approach and the
best design is recommended which maximizes the customer satisfaction value.
Finally, the design is verified by comparing with standard data or developing
prototypes. To design a suitable product (office chair) ergonomically, followings steps
are considered.
3.2.1 Data collection

Data collection consists of two different types of survey in the eastern parts of
India. First, complaints regarding the product (office chair) were collected through
formal and informal interviews. In second survey, the respondents need to answer
the questionnaire consisting of a set of variables, anthropometric sitting and standing
dimensions and a set of design attributes. The study was conducted on different age
group (from 21 to 60 years) of officers and managers working in government and
22
non-government agencies and they used the chairs for a prolonged period every day.
Data are mainly collected from staff of various offices like technical institutions
(private and public), banks (private and public), hospitals (private and public) and
government organization. A cross-sectional survey with random sampling procedure
was conducted. One hundred twenty five responses are obtained. 60% data are
collected from banks, 20% data are collected from technical institutions, and rest of
the data is collected from hospitals and other sources.
Table 3.1 Customer requirements
Sl. Items
No
1 The chair is made for active sitting 1 2 3 4 5
2 The chair is retro looking (designating the style of an earlier 1 2 3 4 5
time)
3 The chair gives pleasant sensation 1 2 3 4 5
4 The chair is immortal (lasting forever) 1 2 3 4 5
5 The chair is chic (stylish and fashionable) 1 2 3 4 5
6 The chair is gaudy one (excessive showy) 1 2 3 4 5
7 The chair is flashy one (too bright intended to get attraction) 1 2 3 4 5
8 The chair is made for active sitting 1 2 3 4 5

9 The chair is very feminine (womanliness) 1 2 3 4 5
10 The chair is very casual one (convenient) 1 2 3 4 5
11 The chair is cute one 1 2 3 4 5
12 The chair is enjoyable 1 2 3 4 5
13 The chair is so cozy (friendly, comfortable) 1 2 3 4 5
14 The chair is untroubled (free from disturbance) 1 2 3 4 5
15 The chair is so cheap 1 2 3 4 5
16 The chair is voluminous 1 2 3 4 5
17 The chair is soft enough 1 2 3 4 5
18 The chair is sturdy enough (strong, solid and thick, unlikely 1 2 3 4 5
to break)
19 The chair is well balanced 1 2 3 4 5
20 The chair is masculine (pertaining to the characteristic of a 1 2 3 4 5
man)
21 The chair is so cool 1 2 3 4 5
22 The chair is most luxurious one 1 2 3 4 5
23 The chair is stylish one 1 2 3 4 5
24 The chair design is contemporary (old but modern feelings) 1 2 3 4 5
25 The chair has a personal recognition 1 2 3 4 5

26 The chair is elegant for you (attractive appearance and 1 2 3 4 5
behaviour)
27 The chair is having distinct features 1 2 3 4 5
28 The chair is simple 1 2 3 4 5
29 The surface of chair is plain 1 2 3 4 5
30 The chair is comfortable 1 2 3 4 5
31 It is an ordinary chair 1 2 3 4 5
32 The chair is flexible 1 2 3 4 5
33 The chair provides headrest 1 2 3 4 5
34 The chair provides spinal curvature support 1 2 3 4 5
23
35 The chair provides footrest 1 2 3 4 5
36 Your leg is comfortable with the chair 1 2 3 4 5
37 The chair is having backrest contour 1 2 3 4 5
38 The armrest is large enough to support your arm 1 2 3 4 5
39 The chair is having a base with wheel 1 2 3 4 5
40 The chair is having lower back support 1 2 3 4 5
Table 3.2 Tangible design attributes

Sl. Backrest Seat pan Arm rest Whole body Others
No
1 Tilt of Backrest(maximum Length of Length of Ratio of seat pan Use of
angle of the backrest in seat pan arm rest and backrest decoration
relation to the seat pan)
2 Width of Backrest Width of seat Width of Width-height ratio Use of
pan armrest of whole body pattern
3 Height of Backrest Thickness of Height of Height of whole Use of
seat pan armrest body cushion
4 Thickness of Backrest Width and Width- Size of whole Use of
length ratio of height body curved
seat pan ratio of lines
armrest
5 Width-Height ratio of Height Number of Number of
Backrest adjustment of controls used colors used
seat pan
Table 3.3 Intangible design attributes

Sl. No Categorical design variables
1 Shape of backrest
2 Material of backrest
3 Colour of backrest
4 Shape of seat pan
5 Material of seat pan
6 Colour of seat pan
7 wheels
8 Low back support
9 Headrest
10 Shape of armrest
11 Material of armrest
12 Colour of armrest
13 Shape of base
Table 3.4 Anthropometric standing dimension (mm)

Sl. No Anthropometric dimension Average
1 Stature(without shoes) 1643.8
2 Eye height 1498.8
3 Cervical height 1425.3
4 Wrist height 810
5 Elbow height 990
6 Waist height 1064
7 Ductylion height 640
8 Ankle height 97
9 Crotch height 876.5
10 Gluteal furrow height 740
24
Table 3.5 Anthropometric sitting dimension (mm)
Sl. No Anthropometric dimension Average
1 Sitting height 806.2
2 Eye height 696.4
3 Cervical height 606.6
4 Upper lumbar height 276.2
5 Elbow rest height 221
6 Popliteal height 459.4
7 Buttock height 455
8 Buttock popliteal length 439
9 body depth 190
10 Shoulder breadth 446
11 Hip breadth seated 361.5
12 Popliteal depth 344.6
Based on existing literature (Han, 2000; Jindo and Hirasago, 1995), informal
discussion with office chair manufacturer and interview with users, a list of adjectives
is used to represent customer’s requirements towards the product as shown in Table
3.1. A total of forty different adjectives indicating customer requirements were
identified. Thirteen different models of office chairs from standard manufacturers are
displayed to the respondents to finalize the tangible and intangible design attributes
of the office chair as listed in Table 3.2 and 3.3 respectively. In fact, the design of
office chairs involves consideration of many tangible and intangible criteria for
reducing fatigue to the users and improving user satisfaction. Ten standing and
twelve sitting anthropometric dimensions of office employees were collected as
shown in Table 3.4 and 3.5 respectively. The customer requirements and design
criteria are suitable to the users having average anthropometric dimension shown in
Tables 3.4 and 3.5. A respondent views the customer requirements taking into
consideration of anthropometric dimensions. The respondent needs to answer in
terms of Likert-type scale from 1 to 5 (1 for strongly disagree and 5 for strongly
agree).
3.2.2 Factor Analysis

The responses obtained through the data collection were tested to examine the
validity and reliability of variable to obtain a statically proven identification of customer
requirements. The validity was tested through factor analysis method using principal
component method following varimax rotation to extract the important dimensions for
model analysis which removes the redundancy and duplication from a set of
correlated variables (Noruzy et al., 2011; Hair et al., 2010).
25
3.2.3 Quality function deployment (QFD)
The reduced customer requirements identified in step 2 must be expressed in
terms of design requirements to provide guidelines for design and manufacturing
engineers while manufacturing the product. QFD plays an important role in this
respect for transforming customer requirements into design characteristics. The main
objective of QFD is to transform customers’ voice (requirement) into design
parameters. House of quality (HOQ) starts with customer requirements, i.e., variables
defined by the customers. These variables are the inputs to the HOQ. Customer
ratings for customer needs are determined by left correlation matrix using equation
3.1. The individual rating of each design requirement is obtained from the central
matrix by using equation 3.2.
 1  n
Customer rating  Z i     B ij Z j (3.1)
 n  1 j1
 1  n 
Design requirements      A ij X j  (3.2)
n  j 
Where, B ij ,denotes the relationship between customer needs, Zi, is the initial
customer rating, A ij , denote the relative importance of i th characteristic with respect to
jth customer’s needs in the relationship matrix, Xj represents the importance of jth
customer needs and n , is the number of customer needs. The refined rating of each
design requirement in the top matrix can be calculated in a similar way as in case of
left matrix. The final ratings of design requirement are normalized by dividing each
rating with the maximum ratings.
3.2.4 Adaptive Neuro Fuzzy Inference System (ANFIS):

Once the design characteristics have been identified in step 3, it is important to
relate the design attributes with customer satisfaction so that best design can be
reached. Since it is not possible to mathematically define the relationship among
design attributes and customer satisfaction due to inherent imprecise nature of
variables, ANFIS can be used to map non-linear relationship for prediction of result
(Jang, 1991, 1993). ANFIS is a combination of two different methodology, i.e., neural
network and fuzzy logic. A neural network can learn from both the data and feedback
without understanding the pattern involved in the data. But, the fuzzy logic models
are easy to comprehend the pattern because they use linguistic terms in the form of
IF-THEN rules. A neural network with their learning capabilities can be used to learn
the fuzzy decision rules; thus creating a hybrid intelligent system. The fuzzy system
26
provides expert knowledge to be used by the neural network. A fuzzy inference
system consists of three components. These are:
(a) rule base, contains a selection of fuzzy rules.
(b) data base, defines the membership functions used in the rules
(c) reasoning mechanism, to carry out the inference procedure on the rules and
given facts.
This combination merges the advantages of fuzzy system and a neural network.
Jang (1991) proposed a combination of a neural network and fuzzy logic popularly
known as called an ANFIS. ANFIS is a fuzzy inference system implemented in the
framework of neural networks. The combination of both artificial neural network and
fuzzy inference system thus improves system performance without interference of
operators. A typical adaptive network shown in Figure 3.1 is a network structure
consisting of a number of nodes connected through directional links. Each node is
characterized by a node function with fixed or adjustable parameters. Learning or
training phase of a neural network is a process to determine parameter values to
sufficiently fit the training data. The basic learning rule method is the back
propagation method, which seeks to minimize some error, usually sum of squared
differences between network’s outputs and desired outputs. Generally, the model
performance is checked by the means of distinct test data, and relatively good fitting
is expected in the testing phase. Considering a first order (Takagi and Sugeno, 1985;
Sugeno and Kang, 1988) fuzzy interface system, a fuzzy model consists of two rules.
Rule 1 : If x is A1 and y is B1 then f1  p1x  q1y  r1
Rule 2 : If x is A2 and y is B2 then f2  p 2 x  q2 y  r2
If f1 and f2 are constants instead of linear equations, we have zero order TSK fuzzy-
model. Node functions in the same layer are of the same function family as described
below. It is to be noted that Oij denotes the output of the ithnode in layer j.
Layer 1: Each node in this layer generates a membership grade of a linguistic label.
For instance, the node function of the ith node might be
1 (3.3)
O ij  A(i x )  bi
2
 x  c  
1   i
 
 a i  
where, x is the input to the node 1 and Ai is the linguistic label (small, large)
associated with this node; and {ai, bi, ci } is the parameter set that changes the
shapes of the membership function. Parameters in this layer are referred to as the
“Premise Parameters”.
27
Layer 2: Each node in this layer calculates the firing strength of each rule via
multiplication
Oi2  w i  A i x   Bi y , i  1, 2 (3.4)
Layer 3: The ith node of this layer calculates the ratio of the ith rule’s firing strength to
the sum of all rule’s firing strengths:
wi
Oi3  w i  ,i  1, 2 (3.5)
w1  w 2
For convenience outputs of this layer will be called normalized firing

strengths.
Layer 4: Every node i in this layer is a squared node with a node function
Oi4  w i fi  w i p i  qi y  ri  (3.6)
Where, w i is the output of layer 3, and is the parameter set. Parameters in this layer
will be referred as “Consequent Parameters “.
Layer 5: The single circle node computes the overall output as the summation of all
incoming signals i.e.
n
O i5  Overall output   w i fi 
w f i i i (3.7)
i w
i
i
Thus, an adaptive network is presented in Figure 3.1 is functionally equivalent to a

fuzzy interface system. The basic learning rule of ANFIS is the back propagation
gradient decent which calculates error signals (defined as the derivative of the
squared error with respect to each nodes output) recursively from the output layer
backward to the input nodes. This learning rule is exactly the same as the beck-
propagation learning rule used in the common feed-forward neural networks by Jang
(1993). From ANFIS architecture (Figure 3.1), it is observed that the given values of
the of premise parameters, the overall output can be expressed as a linear
combination of the consequent parameters. Based on this observation, a hybrid
learning rule is employed here, which combines a gradient decent and the least
squares method to find a feasible of antecedent and consequent parameters. The
details of the hybrid rule are given by Jang (1993) where it is also claimed to be
significantly faster than the classical back propagation method. From the ANFIS
architecture shown in Figure 3.1, we observe that when the values of the premise
parameters are fixed and the overall output can be expressed as a linear
combination. The output f can be rewritten as:
w1 w1
f f1  f2
w1  w 2 w1  w 2
28
= w f1  w f2
         
= w x p1  w y q1  w1 r1  w 2 x p 2  w 2 y q2  w 2  r2 (3.8)
which is linear in the consequent parameters p1, q1, r1, p2, q2, r2. Therefore, the hybrid
learning algorithm developed can be applied directly. More specifically, in the forward
pass of the hybrid learning algorithm, node outputs go forward until layer 4 and the
consequent parameters are identified by the least squares method. In the backward
pass, the error signal propagates backward and the premise parameters are updated
by gradient descent. As mentioned the consequent parameters thus identified are
optimal under the condition that the premise parameters are fixed. Accordingly, the
hybrid approach converges much faster since it reduces the dimension of the search
space of the original back-propagation method. For this network created fixes the
membership functions and adapt only the consequent part; then ANFIS can be
viewed as a functional-linked network where the enhanced representation, which take
advantage of human knowledge and express more insight. By fine-tuning the
membership functions, we actually make this enhanced representation.
Figure 3.1 A typical architecture of ANFIS structure

The data collected on office chairs through QFD model analysis are normalized.
The selected normalized design elements are considered as inputs for ANFIS system
and the overall customer satisfaction for chair is considered as output. Total
experimental data set is divided into training and testing data set. A total of 80
datasets are used in ANFIS model. Sixty datasets are considered as training and
twenty datasets are considered under testing. During training, a five layered ANFIS
structure is constructed with one input, three hidden and one output. The Gaussian
type of membership function (gaussmf) is used for input and linear type function is
used for output. The number of correct outputs is noted till the error is minimized.
29
3.2.5 Design of Experiment:
DOE is a cost effective statistical method used to optimize the response of a
process under the influence of multiple factors. The information on factor and
interaction effects on response can be estimated with less number of experimental
runs. In fact, it involves multiple factors during experimentation to find out optimum
treatments for best performance. The factors with different levels (high and low) are
the input parameters which affect the output performance. The best combination is
selected out of a number of combinations of different levels of different factors. DOE
extensively uses full factorial and fractional factorial experiments to optimize process
parameters. Since fractional factorial experiments requires less number of
experiments compared to full factorial design, Taguchi proposes simplification and
standardization of fractional factorial designs to estimate main effect at different
levels of factors and optimize the parameters (Wu and Hamada, 2000).
3.3 Results and discussions

One hundred twenty five data are collected from respondents through cross-
sectional survey on 40 items as shown in Table 3.1 to consolidate the customer
requirement on office chair. The survey data are subjected to factor analysis to
eliminate redundancy of data. Factor analysis has been carried out SPSS 14.0.
Twenty two customer requirements were loaded on three factors showing factor
loading score more than 0.7.The items exhibiting factor loading score of 0.7
(threshold value) are not considered further. Total variance explained by three factors
was found to be 78.5% which is acceptable value for principal component with
varimax rotated factor loading procedure. Ten items were loaded under factor 1, five
items under factor 2 and seven items under factor 3 (Table 3.6). Factors extracted
from analysis are named as comfortness (factor 1), balance (factor 2) and
luxuriousness (Factor 3). Cronbach’s alpha (α) has been used to assess the internal
consistency of the scale. The value of alpha for all dimensions is 0.702, which is just
the acceptable value of 0.70 for demonstrating internal consistency of the established
scale. The values of α obtained are 0.878, 0.933, and 0.939 for factors 1, 2, and 3
respectively. The Kaiser-Meyer-Olkin (KMO > 0.6) and Bartlett’s test of sphericity (p <
0.05) statistics are used to test empirically whether the data were likely to factor well.
The value of KMO was found to be 0.665; hence, it can be concluded that the matrix
did not suffer from multi collinearity or singularity. The result of Bartlett’s test of
sphericity shows that it is highly significant (sig. = 0.000) which indicates that the
factor analysis is correct and suitable for testing multidimensionality.
30
Table 3.6 Factor Analysis for all items (0.702)
Dimensions Variables Factor Loadings
Comfortness(0.878) 1.Active sitting 0.729
2.Cute 0.833
3.Enjoyable 0.850
4.Friendly 0.820
5.Untroubled and free 0.719
from disturbance
6.Masculine 0.719
7.Having plain surface 0.825
8.Comfortable 0.882
9.Flexible 0.816
10.Having lower 0.778
back support
Balance(0.933) 1.Lasting forever 0.885
2.Soft enough 0.865
3.Sturdy enough 0.875
4.Well Balanced 0.843
5.Having Footrest 0.782
Luxuriousness(0.939) 1.Excess showy 0.715
2.Too bright, intended 0.753
to get attraction
3.Refind one 0.720
4.Luxurious one 0.775
5.Stylish one 0.742
6.Having personal 0.744
recognition 0.808
7.Elegant for you
The customer requirements are classified into three factors such as comfortness,
balance and luxuriousness. The items under each factor are expressed by the
customers in a vague sense. To provide guidelines for the manufacturing, the vague
items under each factor needs to be converted into design attributes. QFD being a
suitable method of converting vague customer requirements into tangible design
attributes, it is used for establishing the relationship among customer requirement
and design attributes through experts’ opinion and brainstorming sessions. Three
different QFD models named as QFD model 1 (comfortness), QFD model 2 (balance)
and QFD model 3 (luxuriousness) are used for correlating customer requirement with
design attributes. Ten, five and seven items (customer requirements) are considered
under QFD models 1, 2 and 3 respectively (Table 3.6). The design attributes
extracted from the experts for three models are shown in Table 3.7. The design
attributes considered are nine, seven and nine for models 1, 2, and 3 respectively.
31
Table 3.7 Design attributes for three different QFD models
QFD model 1(comfortness) QFD model 2(balance) QFD model3(luxuriousness)
Tilt of backrest Width-height ratio of Seat adjustment range
backrest
Number of controls Base material Use of pattern
Width of backrest Size of base wheel Use of curved line
Depth of seat pan Width-height ratio of seat Use of cushion
pan
Height of armrest from the Width-height ratio of whole Use of colour
floor body
Overall height Thickness of seat pan Shape of backrest
Overall width Width-height ratio of armrest Use of decoration
Low back support Shape of seat pan
Width of seat pan Backrest height
Initial rating of customer requirements for each model is derived using a 1 to 10

scale as shown for the case of model 1, 2, 3 in Figure 3.2, 3.3 and 3.4. The customer
ratings for each customer requirement were obtained from left correlation matrix
using equation (1) and initial design requirements is obtained from central matrix
using equation (2). The correlation of customer requirements (left matrix), design
requirements (top matrix) and customer requirements with design attributes (central
matrix) are extracted from the experts using scale of 0.8, 0.6, 0.4 and 0.2 for
designating relationship ‘strong’, ‘moderate’, ‘weak’, and ‘very weak’ respectively.
Finally, initial design requirements and with correlation values shown in top matrix are
used in equation (1) to obtain final design ratings. The normalized refined rating of
design attributes are obtained by dividing each rating with the maximum available
design requirement rating. In Figure 3.2, the QFD model 1 (comfortness) is shown.
From the normalized refined rating for design attributes, ‘Tilt of backrest’ has the most
prioritized followed by ‘Number of controls’ (Table 3.8). Finally, four design attributes
such as tilt of back rest, number of controls, overall width and overall height are
considered out of nine design attributes having normalized refined rating value of
0.85 (threshold). Similarly, other two models have been developed. For QFD model 2
(balance), four design attributes such as width-height ratio of backrest, width-height
ratio of seat pan, width-height ratio of whole body, and width-height ratio of
armrest(Table 3.9) exhibiting normalized refined rating value of 0.80 (threshold) and
above have been considered. Similarly, five design attributes such as seat
adjustment range, use of pattern, use of cushion, use of decoration, and backrest
height showing (Table 3.10) normalized refined rating value of 0.90 (threshold) and
above have been considered for QFD model 3 (luxuriousness).
32
Table 3.8 Ranking of Design attributes for model-1(comfortness)
No Design requirement Initial Design Revised Normalized Rank

requirement Design refined rating
Rating requirement
Rating
1 Tilt of Back rest 5.994 8.642 1.000 1
2 Overall height 4.778 7.153 0.856 4
3 Overall Width 5.136 8.045 0.931 3
4 Height of armrest 4.510 7.400 0.828 5
5 Low back support 3.484 5.425 0.628 9
6 Length of seat pan 3.479 6.149 0.711 8
7 Size of base wheel 3.824 6.244 0.722 7
8 Number of controls 5.659 8.341 0.965 2
9 Height of backrest 4.005 6.946 0.804 6
Table 3.9 Ranking of Design attributes for model-2(balance)
No QFD model 2(balance) Initial Revised Normalized Rank

Design Design refined
requirement requirement rating
Rating Rating
1 Width-height ratio of backrest 6.624 10.450 0.962 3
2 Base material 3.662 5.830 0.537 7
3 Width-height ratio of seat pan 7.486 10.861 1.000 1
4 Thickness of seat pan 3.642 7.433 0.684 6
5 Width-height ratio of whole 6.840 10.515 0.968 2
body
6 Width-height ratio of armrest 6.440 10.040 0.924 4
7 Size of base wheel 4.488 7.957 0.733 5
Table 3.10 Ranking of Design attributes for model-3(luxuriousness)
No Design requirement Initial Design Revised Normalized Rank

requirement Design refined
Rating requirement rating
Rating
1 Seat adjustment range 6.516 10.016 0.975 3
2 Use of pattern 6.909 10.258 0.998 2
3 Use of cushion 5.383 8.255 0.803 5
4 Use of curved line 4.396 7.284 0.709 6
5 Use of colour 4.480 7.127 0.694 7
6 Shape of backrest 3.966 6.318 0.615 9
7 Use of decoration 6.356 9.971 0.970 4
8 Backrest height 6.816 10.276 1.000 1
9 Shape of seat pan 4.482 6.696 0.652 8
33
Figure 3.2 QFD model for comfortness
34
Figure 3.3 QFD model for balance
35
Figure 3.4 QFD model for luxuriousness
One hundred twenty five survey data collected from the respondents over 13
types of chairs regarding satisfaction level of the users are used in the ANFIS model
to develop the relationship between design attributes and customer satisfaction.
Three ANFIS models have been developed to relate QFD model 1 (comfortness),
QFD model 2 (balance) and QFD model 3 (luxuriousness). The inputs to each ANFIS
model is the design attributes related to the type of QFD model obtained over 13
types of chairs during survey (basically these are chair dimensions or tangible design
attributes shown in Table 3.2). The output of each model is nothing but the sum of the
customer requirements. For example, ANFIS model for QFD model 1 (comfortness)
36
treats four design attributes such as tilt of back rest, number of controls, overall width
and overall height as inputs and sum of response value of ten customer requirements
such as active sitting, cute, enjoyable, friendly, untroubled and free from disturbance,
masculine, having plain surface, comfortable, flexible, and having lower back support
as outputs. In each model, 94 data (75%) are used for training and 31 data (25%)
used for testing. The data in each model is normalized by diving corresponding
maximum value. The ANFIS architecture for three QFD models: (a) comfortness, (b)
balance, (c) luxuriousness is shown in Figure 3.5. Input membership function is
described with Gaussian membership function. Hybrid learning algorithm is used and
ANFIS model is run till the error is minimized. Error is minimized in three epochs
during training. Then, testing of data is carried out. The pattern of variation of actual
and predicted response is shown for training and testing dataset for QFD model 1
(comfortness). Figures 3.6 and 3.7 shows that actual (blue dot) and predicted (red
dot) values for three models which are uniformly distributed respectively for training
and testing data.
(a) comfortness
37
(b) balance
(c) luxuriousness
Figure 3.5 ANFIS model structure: (a) comfortness, (b) balance, (c) luxuriousness
38
(a)
(b)
39
(c)
Figure 3.6 Distribution of predicted and actual response training: (a) comfortness, (b)
balance, (c) luxuriousness
(a) comfortness
40
(b) balance
(c)luxuriousness
Figure 3.7 Distribution of predicted and actual response testing: (a) luxuriousness, (b)
balance, (c) luxuriousness
41
The surface plot shown in Figure 3.8 indicates that the total landscape of decision
space is covered by the ANFIS model for QFD model 1 (comfortness).For the model
1(comfortness), Input 1 indicates ‘tilt of backrest’ and input 2 indicates ‘number of
controls’. Similarly for model 2 (balance), input 1 is ‘width-height ratio of seat pan’ and
input 2 is ‘width-height ratio of whole body’ and for model 3(luxuriousness), input 1 is
‘backrest height’ and input 2 is ‘use of pattern’.
(a) comfortness
(b) balance
42
(c) luxuriousness
Figure 3.8 Surface plots: (a) comfortness (b) balance(c) luxuriousness
The residual analysis is carried out for the predicted values of the model by
calculating the difference of actual and predicted values for training and testing data.
The residual plots for three factors are depicted in Figure 3.9. It is observed that the
residuals for three factors are distributed uniformly along the centre line. The absolute
percentage relative error in training phase is 0.00039 and in testing phase 0.122164.
The residuals are distributed normally when tested with Anderson-Darling test
statistic. Similar procedure is adopted to predict the response for QFD model 2
(balance) and 3 (luxuriousness). The absolute percentage relative error for QFD
model 2 (balance) is 0.025632 (training) and 0.063265 (testing). Similarly, absolute
percentage relative error for QFD model 3 (luxuriousness) is 0.068057 (training) and
0.124575 (testing).Therefore, it can be stated that prediction of customer satisfaction
can be made with ANFIS accurately.
(a) comfortness
43
(b) balance
(c) luxuriousness
Figure 3.9 Residual plots: (a) comfortness (b) balance(c) luxuriousness
The Anderson-Darling test (AD Test) is also carried out in order to compare the fit
of an observed cumulative distribution function to an expected cumulative distribution
function. Smaller the AD value, greater is the evidence that the data fit to the normal
distribution. The test results are shown in Figure 3.10 for respective factors
(comfortness, balance, luxuriousness) standardized residue. The following figures
suggest that all the data are normally distributed for the data obtained from ANFIS
model. Similarly, the normal probability plot of residual for testing data of three
models is shown in Figure. Since p-value of the normality plots is found to be above
0.05, it signifies that residue follows normal distribution.
44
(a) comfortness
(b) balance
45
(c) luxuriousness
Figure 3.10 Normal probability curve of residual at 95% confidence
level:(a)comfortness(b)balance(c)luxuriousness
Table 3.11 Levels of various designs attributes (parameters)
Symbol control Actual levels Coded levels
parameters
Low Medium High Maximum Low Medium High
A Use of pattern No Yes 0.000 1.000
B Use of cushion No Yes 0.000 1.000
C Use of decoration No Yes 0.000 1.000
D Height 89.7 101.6 112.3 132.6 0.677 0.767 0.847
adjustment(mm)
E Overall height of 370.5 426.0 528.0 584 0.635 0.730 0.921
backrest(mm)
F Tilt of backrest(in 18 19 26 32 0.563 0.594 0.813
degree)
G Number of No One Two Two 0.000 0.500 1.000
controls
H Total height(mm) 990 1,110 1,245 1,245 0.795 0.892 1.000
I Total width(mm) 535 620 838 838 0.638 0.740 1.000
J Width-height ratio 0.712 0.833 0.935 1,066 0.668 0.782 0.877
of backrest
K Width-Length 0.941 0.992 1.024 1,128 0.834 0.879 0.907
ratio of seat pan
L Width-height ratio 0.200 0.486 0.638 0.854 0.234 0.569 0.747
of whole body
M Width-height ratio 0.228 0.319 0.55 0.55 0.415 0.580 1.000
of armrest
46

Table 3.12 Experimental design using orthogonal array of l36 2 3  310 
Thirteen important design attributes meeting all the three requirements such as
comfortness, balance and luxuriousness of the customers (nine design attributes for
comfortness, four design attributes from balance and five design attributes for
luxuriousness) are treated as various factors for the design of office chair. The
relationship of these attributes with customer satisfaction is established through
ANFIS model. In order to search the best design, a large number of scenarios were
generated using DOE approach. The criteria for best design parameter are based on
‘larger-the-better’ type, i.e., maximizing customer satisfaction level. The design
attributes (parameters) are use of pattern, use of cushion, use of decoration, number
of control used, range of height adjustment, backrest height, tilt of backrest, overall
height, overall width, width-height ratio of backrest, width-length ratio of seat pan,
width-height ratio of whole body and width-height ratio of armrest. Out of 13
parameters (design attributes), three parameters, each having two levels (low and
high) and ten parameters, each having three levels (low, medium, high) are
considered. This requires a total of (23 * 310) = 472,392 experiments. But Taguchi’s
mixed level experiments can produce same information using 36 experiments. Each
parametric level is divided by the maximum value of design attribute (parameter) to
set the levels. Because of frequency of occurrence maximum parametric values is
comparatively less as observed in the survey. Therefore, frequently occurring
47
parametric values are used to set the levels as shown in Table 3.9. Three ANFIS
models (for QFD models 1, 2 and 3) were run considering respective design
attributes to obtain the response (customer satisfaction). The customer satisfaction
values obtained from three models are summed to provide the response for DOE
scenario generation assuming equal weightage for customer requirement factors
such as comfortness, balance and luxuriousness. The experimental layout along with
the responses is shown in Table 3.10. The main effect plot shown in Figure 3.11 finds
the optimal setting of design attributes as A1 B1 C0 D0.767 E0.635 F0.594 G0.5 H0.892 I0.638
J0.877 K0.879 L0.569 M0.415.
Figure 3.11 Effect of control design parameters on response
The values of optimal design attributes of office chair are shown in Table 3.11.
These values are compared with office chair dimensions of BIS. However, BIS does
not provide values of all the attributes considered in this work. Therefore, a few
dimensions (design attributes) are selected to verify the design. It can be observed
from the Table 11 that four design attributes such as height range adjustment (D),
total height of backrest (E), whole body height (H), and whole body width (I) are
within the limits. However, height range adjustment is slightly (1 to 2 millimeters)
above the BIS limit. Four BIS elements such as width of the seat, depth of the seat,
height of seat pan from the floor (higher range), and height of seat pan from the floor
(lower range) are considered for comparing the design attribute values so obtained
(Table 3.12). Width of the seat is obtained by deducting sum of armrest width and
48
clearance from the overall seat width (arm rest width is obtained as 30 mm from the
survey data and arm rest width both sides is 60 mm whereas a clearance of 3.5 mm
is assumed). The width of seat is calculated as 471.5 mm which is more than the BIS
specification of 450 mm. Depth of the seat is obtained by dividing width of seat by
optimal width-depth ratio of seat pan. The calculated depth of seat is 475 mm which
is above the minimum BIS limit. Height of seat pan from the floor (higher range) is
calculated by subtracting total height of backrest from whole body height. The value
is 452 mm which is below the maximum limit of 500 specified by BIS. Height of seat
pan from the floor (lower range) is calculated by subtracting height range adjustment
from height of seat pan from the floor (higher range). The value is calculated as 350
mm which is below the minimum BIS specification of 400 mm.
Table 3.13 Comparison of optimum parameter with Bureau of Indian Standard (BIS)
Sl. No Design attributes Optimal values BIS

of design specifications(mm)
attributes
A Use of pattern Yes
B Use of cushion Yes
C Use of decoration No
D Height range adjustment(mm) 101.6 100
E Total height of backrest(in 538.0 250(minimum)
degree)
F Tilt of backrest 190
G Number of controls 1
H Whole body height(mm) 990.0 785.0(minimum)
I Whole body width(mm) 535.0 535.0(minimum)
J Width-height ratio of backrest 0.9347
K Width-height ratio of seat pan 0.992
L Width-height ratio of whole body 0.486
M Width-height ratio of arm rest 0.228
By considering these dimensions, a prototype of office chair is made using Auto
CAD Version 10 as shown in Figure 3.8 and3.9.
49
Table 3.14 Dimensions of prototype (excluding known control parameter) for
comparison
BIS elements Calculation from Optimal BIS
design attributes values(mm) specifications(in
mm)
Width of the seat Whole width of seat 535.0- 450(minimum)
including armrest- (60+3.5)=471.5
(width of armrest
clearance)
Depth of seat Width of seat  width 471.5  400(minimum)
depth ratio of seat pan 0.992=475.0
Height of seat pan Whole body height- 0.992- 500(maximum)
from the floor(higher Total height of 538.0=452.0
range) backrest
Height of seat pan Height of seat pan 452.0- 400(minimum)
from the floor(lower from the floor(higher 101.6=350.4
range) range)-height range
adjustment
50
Figure 3.12 (a) Orthographic Projection (All dimensions are in mm)
51
Figure 3.12 (b) Prototype of office chair with optimized design parameter
(All dimensions are in mm)
3.4 Conclusions
Ergonomically designed industrial or office work environment considering both
psychological and physical needs of employees helps to reduce fatigue when the
employees continue prolonged work. Use of ergonomically designed equipment not
only increases job satisfaction but also injury level can be prevented at the work
place; hence enhance employees’ efficiency and productivity. Since the design
includes both subjective and objective criteria, it is not easy to design a product that
improves user satisfaction. Therefore, an integrated approach using statistical and
artificial intelligence techniques has been proposed in this article. The approach is
described with the help of an office chair design. The user/customer requirements
have been extracted through a cross-sectional survey. Factor analysis has been
52
carried out on data to eliminate redundancy. The customer requirements are mapped
to design attributes using QFD. A functional relationship has been developed among
design attributes with customer satisfaction using adaptive neuro-fuzzy system.
Finally, a Taguchi robust design approach is adopted to generate various scenario of
office chairs having varied design attributes. Finally, the design that maximizes
customer satisfaction has been chosen. The optimal design so obtained is compared
with design specifications laid down in BIS. The proposed design satisfies most of
comparable design elements of BIS. The variations are attributed to localization of
sample data. The approach is quite general and can be adopted in any design.
53
CHAPTER 4
A NUMERICAL APPROACH
FOR ERGONOMIC DESIGN
4.1 Introduction
Pressure ulcer is the localized area of tissue degeneration in sub-dermal tissue as
a result of prolonged continuous mechanical load (National Pressure Ulcer Advisory
Panel, 1989). External mechanical load (weight of the body) always induces a
mechanical deformation in soft tissue (Chow and Odell, 1978). Excess pressure for a
long time restricts the blood vessels resulting in the formation of tissue ischemia and
ultimately tissue necrosis (Crenshaw and Vistnes, 1987). When a larger pressure is
applied to soft tissue, it decreases the time of causing cell death and the tissue
damage starts due to impaired capillary profusion giving rise to hypoxia (Kosiak et al.,
1958). It has been indicated that the pressure sore mostly occurs at the lower part of
body i.e. 43% at the sacrum and 5% at the ischial tuberocity (bony part). To lower the
stress distribution, either the intensity of load or sitting time duration is to be reduced
but at the same time the work at an office environment must not be compromised
(Peterson,1976). Therefore, design modification of the product (chair seat cushion) or
choosing the product with suitable properties should be emphasized to achieve this
goal. Sitting comfort for a long time can decrease the rate of cell death. Seat cushion
properties can be useful in reducing the deformation of tissues. Polyurethane foam
with different properties can be used to investigate stress distribution at ischial
tuberocity. Recently, number of studies have reported the effect of foam density,
foam compressibility, strain rate and energy absorption of polymeric foam under
uniaxial loading. Polyurethane foam under polymeric foam category is mostly used
for seat cushion because it can undergo large compressive deformation and not only
absorb but also dissipate considerable amount of energy under loading in
comparison to solid specimen of equal volume (Avelle, 2001). Polyurethane foam
exhibits viscoelastic behavior which depends upon the time scale of loading and
temperature of material (Gibson, 2012; Lakes, 1999; Briody et al., 2012; Mills, 2007;
Schrodt et al., 2005).
As the tissue lying below the bony part exhibits maximum stress under loading
and undergoes tissue deformation, the present study focuses on the stress
distribution at the tissue near bony prominence (ischial tuberocity) through numerical
analysis by changing cushion properties. In this context, the mechanical condition of
seat is considered by changing seat material properties and parameters to study
change in stress distribution at ischial tuberocity. A rigid seat-buttock model is also
considered for validation purpose. As the change in cushion properties are alone not
sufficient for reduction in stress distribution at ischial tuberocity, further analysis with
different thickness of cushion and loading angles has also been carried out. The
objective of the present work is to develop a simple two dimensional finite element
54
buttock-seat model with various material parameters, thicknesses of seat with
different loading angle to predict stress at ischial tuberocity in order to provide
guidelines to reduce occurrence of pressure ulcer.
4.2 Model descriptions
A volunteer weighing fifty five kilogram is contacted to obtain image of the buttock
using magnetic resonance imaging (MRI) technique. The volunteer is scanned
against a weight bearing posture sitting on MR compatible plastic chair with erect
back rest and buttock support area 400 mm wide. Markers on the chairs are used to
maintain the buttock position for the cushion by aligning the midline of body with a
frontal marker on chair (Shabshin et al., 2010). Data on buttock of seated human
including fats, muscles and ischial tuberocity is extracted from a double donut 0.5T
open MR system for developing a 2D model representing buttock (Linder-Ganz et
al.,2007; Linder-Ganz et al.,2008). Using the data from MRI, the position of ischial
tuberocity, muscle and fats are traced and the boundary is located by maintaining the
distance between ischial tuberocity and skin (Tang et al., 2010). A dimensional
buttock model is made for exporting to simulation software. A non linear FE analysis
using ANSYS 10.0 is developed to obtain the maximum stress distribution at ischial
tuberocity due to the interaction of different seat cushion and buttock. Two different
models comprising of human soft tissue-rigid seat (model I) and human soft tissue-
soft polyurethane foam cushion (model II) are considered. The model includes a soft
human tissue, ischial tuberocity (a bony part) and the seat as shown in Figure 4.1.
Buttock-seat model is assumed to be two dimensional axisymmetric finite element
model with an upright posture. Ischial tuberocity is assumed to be a circle of radius
ten millimeters. The seat cushion has been modeled by a rectangular flat surface
having thickness of 80 mm and area of 450×450 mm2 (following standard office chair
width of Bureau of Indian Standards specification). In order to analyze the structure,
the model is divided into small sub-domains (elements) and the equations are
discretized and solved within each of the sub-domains (elements).
55
Figure 4.1 Model of seat cushion and buttock (soft tissue)
As seen from Figure 4.2, x-y plane is considered to present the meshed
axisymmetric model of buttock as well as seat with boundary conditions. Bottom of
the seat is fixed from all directions. Contact interface is determined to make an
interaction between seat and buttock. Load is applied due to the weight of the upper
part of the body. The load carried by the soft tissue is 22kg which is the half of the
upper part of an average weighed human being (Tang et al., 2010). The components
of both soft tissue and seat model are meshed with four-node first order quadrilateral
finite elements. The simulation is conducted over a period of three hours to study the
effect of time of loading on soft tissue. Displacement for all nodes along the
axisymmetric line is fixed. The seat-buttock model is meshed with 1194 four-node
quadrilateral solid finite elements. Plane 182 element type is used for simulation of
both seat and buttock. An element is defined by four nodes and each node carries
two degrees of freedom (translation along x and y directions). To model the
interaction between human buttock and seat cushion, contact element CONTA 171 is
used. A total of 100 steps are considered for each simulation run.
56
Figure 4.2 Finite element model of seat cushion and buttock (soft tissue)
The stress distribution throughout meshed model is studied by varying the

material properties of seat cushion, angle of loading (sitting posture) and cushion
thickness. The angle of loading ranges from 00 to 300. The thickness of cushion
ranges from 60 mm to 80 mm.
4.2.1 Material properties for seat
Since polyurethane foam falls under elastomers category, hyper-elasticity is used
to describe the material properties. Odgen hyper-elastic model based on stretch ratio
is considered to describe the current state of deformation in soft cushion (Odgen,
1997).The governing equation used for the analysis is a strain energy function given
as follows.

2 i  ˆ i ˆ i ˆ i
  

N
1 elii
U 2  1
  2  3  3  J  1 (4.1)
i1 i  i 
where N is the order of fitting, μi , αi and βi are the temperature dependent material
parameters to be determined, J el and J th are the elastic volumetric deformation and
thermal volumetric deformation respectively, ˆ 1 , ˆ 2 , ˆ 3 are the principal stretch ratios
which provide a measure of deformation.

All the polyurethane foams exhibit some visco-elastic properties with deformation
depending upon the load, time, and temperature. Stress relaxation is an important
factor for analysis as the deformation exists even after the removal of the stress.
Visco-elastic behavior occurs in the process of prolonged sitting and the behavior is
defined in terms of time based Prony series model. The governing equation is given
as:
57
N  t

G( t )  G0   Gi 1  e i  (4.2)
 
i1
 
where  i (relaxation time),Gi (relaxation modulus) and G0 (instantaneous shear

modulus) are the material dependent parameters determined by relaxation test and N
is the order of Prony series.
Three different types of seat materials are considered for analysis. For a rigid seat
(model I), low carbon steel with elastic modulus of 210 GPa and Poisson’s ratio of 0.3
is considered. The density (ρ) of the rigid seat is considered as 100kg/m3. Model II
allows the simulation of polyurethane foam cushion with two different material
stiffness and density. Due to the rigidity of the bony part (ischial tuberocity), it allows
only linear elastic material properties and having stiffness of E=80GPa and density,
 =1600kg/m3. The model II undergoes large elastic deformation with approximately
70% deformation for both the polyurethane material. One of the polyurethane foam
considered as SAF 6060 having modulus of elasticity of E=200 kPa and density  =60
kg/m3. Material parameters for SAF 6060 cushion is highlighted in Table 4.1. A strain
energy function of second order under uniaxial compression test was implemented to
find out the material properties of soft foam (Schrodt,2005).The test indicates that
material properties of the foam depend upon the temperature and humidity. Similarly,
by choosing the order of Prony series and performing a curve fitting from the
experimental results of relaxation test, the visco-elastic parameters can be estimated
(Gru-jicic, 2009). Table 4.2 shows visco-elastic material parameters under time based
Prony series model of second order.
Table 4.1 Material properties for Ogden hyper-foam (SAF 6060) for soft cushion
(E=200 kPa,  = 60 kg/m3)
1 (MPa) 1 1  2 (MPa) 2 2
−2 2 −1 −2 2
0.481 × 10 0.198 × 10 0.145 × 10 0.36 × 10 0.198 × 10 0.65 × 10−2
Table 4.2 Coefficients of Prony series parameters for cushion exhibiting

viscoelasticity
N G (i) (sec)
1 0.3003 0.010014
2 0.1997 0.1002
The second polyurethane foam considered in the present analysis also belongs to
polyurethane foam category but having elastic modulus of E=20 kPa and density of
 = 40 kg/m3. The data of material properties is obtained from uniaxial compression
58
test (Briody, 2011). The Ogden hyper-foam model of second order is shown in Table
4.3. A set of Prony series material parameters of third order from relaxation test is
defined in Table 4.4 (Briody, 2011).
Table 4.3 Material properties for Ogden hyper-foam for soft cushion (E=20 kPa,  =
40 kg/m3)
1 (MPa) 1 1  2 (MPa) 2 2
1 5
0.44185  10 21.4556 0 .37050  10 -6.8900 0
Table 4.4 Coefficients of Prony series parameters for cushion exhibiting

viscoelasticity
N G (i)  (sec)
1 0.0973 0.30639
2 0.1740 11.21
3 0.1290 1011
4.2.2 Material properties for human soft tissue

The buttock part of the human soft tissue is modeled with visco-hyper-elastic
model. This hyper-elastic behavior of soft tissue is represented by polynomial strain
energy potential function (U) based on strain invariant which shows the nonlinear,
incompressible, isotropic, hyper-elastic polynomial behavior.
    
N N
1 el
 Cij I1  3 I2  3  
i j 2i
U J 1 (4.3)
i j1 i1 D i
where U is the strain energy potential, J el is the elastic volumetric deformation, I1̅ and
I2̅ are the principal invariants which are independent of coordinate system used to
measure the strain. N is the order of fitting and C ij and D i are the material
parameters. C ij describes the shear behaviour of the material and D i denotes the
compressibility. In this work, second order hyper-elastic parameters of tissue are

considered as shown in Table 4.5 (Tang et al., 2010). The density of soft tissue is
considered as 1000 kg/m3 (Pennestrì, 2005).
Table 4.5 Material parameters for hyperelastic material to define the soft human
tissue
C10 C01 C20 C11 C02 D1 D2
0.08556 -0.05841 0.039 -0.02319 0.00851 3.65273 0
Viscoelastic parameter for soft tissue are considered as shear modulus, G1=0.5, Bulk
modulus, K1=0.5 and relaxation time,  1 =0.8 secs which are taken from the work of
Tang and Tsui,2006.
59
Model I as shown in Figure 4.1 is simulated for three hours using ANSYS 10.0
and the von Mises stress distribution at ischial tuberocity after each half an hour
interval is noticed to quantify the stresses for erect sitting posture. Figure 4.3 shows
the variation of stress at ischial tuberocity for a rigid seat (Model I) at each half an
hour for different seat thickness. Figure 4.4 shows the red patches depicting high
stress region which is most probably the damage area for seat thickness of 80 mm.
Both the Figures 4.3 and 4.4 indicate that the von Mises stress at ischial tuberocity
goes on increasing with sitting time. However, the stress becomes constant after an
interval of one and a half hours for seat thickness of 60 and 70 mm and two hours for
seat thickness of 80 mm (Figure 3). It is evident from Figure 4.4 that the area of high
stress region increases with time. Experimental studies report that tissue stiffening
occurs after one hour of continuous loading of 32kPa (Linder-Ganz et al., 2006). The
stress developed in human tissue within one hour of sitting is sufficient for causing
cell death. After one hour of sitting, tissues have been stiffened and hence stresses
on the tissue may not increase substantially. However, the intensity of load for
stipulated time duration significantly influences cell damage. For example, a pressure
of 11.5kPa for 360 minute of loading can cause the same damage as pressure of
35kPa for 15 minutes (Linder-Ganz et al., 2007). Figure 4.4(a) shows a maximum von
Mises stress of 37193Pa at ischial tuberocity after sitting for half an hour. The result
obtained through the present work is comparable with experimental value (26.7 kPa)
obtained in Verver et al. (2004).
Rigid seat (density=100kg/m3) at an angle of tilting 00
39500
von Mises stress in Pascal
39000
38500
38000
37500
37000
36500
36000
1800 3600 5400 7200 9000 10800
Time duration in sec
Figure 4.3 Variation of von Mises stress with increase in time for different thickness
of seat
60
Figure 4.4 (a) Finite element model for rigid seat with magnification of muscle region
subjected to high stress
Figure 4.4 (b) Stress distribution at ischial tuberocity for erect immobilized continuous
sitting on rigid seat of thickness of 80mm for 1800 sec (von Mises
stress=37193Pa High stress area=11.20 mm2)
61
Figure 4.4 (c) Stress distribution at ischial tuberocity for erect immobilized continuous
sitting on rigid seat of thickness of 80 mm for 5400 sec (von Mises
stress=37997 Pa High stress area=17.02 mm2)
Figure 4.4 (d) Stress distribution at ischial tuberocity for erect immobilized continuous
sitting for rigid seat of thickness of 80 mm for 9000 sec (von Mises
stress=38442 Pa High stress area=23.36 mm2)
4.3.1 Model validation

The model is validated using rigid seat as described above. However, sitting on
rigid seat is not a practical condition in an office environment. Soft cushion is used to
provide a comfortable working environment. Hence, further analysis is carried out on
soft cushion considering Model II with different cushion properties and cushion
62
thickness. Experimental studies show that mild tissue stiffness generally occurs just
after ninety minutes of continuous loading on a rigid seat due to tissue dehydration
(Gefen et al., 2005). Therefore, a soft cushion should be considered in order to avoid
the stiffening of soft muscle tissues in a short interval of time. As the visco-elastic
materials have the capacity to absorb the energy on impact of load (Ferguson-Pell,
1990), soft cushion made of polyurethane foam with two different set of visco-hyper-
elastic cushion properties is considered here to estimate the effect of cushion
properties in predicting stress distribution in muscle tissue. The Young’s modulli and
density for these two cushion type are E1=200kPa, E2= 20kPa, and 1  60kg / m
3
 2  40kg / m 3 having cushion thickness of 80 mm.

As shown in Figure 4.5 (a and b), for both polyurethane cushion properties of
model II, comparatively less stress is developed at ischial tuberocity than that of rigid
seat (Figure 4.3). The pattern of increase in stress is same as that of rigid seat for
one and a half an hours. After one and half hour, the stress go on increasing at
slower rate up to three hours instead of remaining constant as in case of rigid seat.
This phenomenon indicates that the tissue muscles take comparatively longer time to
become stiffened when interact with soft cushion than that of rigid seat.
Soft cushion (E1=200kPa and ρ1=60kg/m3) at an angle of tilting 00
25000
24500
24000
23500
23000
22500
22000
21500
21000
20500
1800 3600 5400 7200 9000 10800 12600
(a)
63
Soft cushion (E2=20kPa and ρ2=40kg/m3) at an angle of tilting 00
25000

24000
23000
22000
21000
20000
19000
18000
1800 3600 5400 7200 9000 10800 12600
(b)
Figure 4.5 Variation of von Mises stress over time for soft cushion of (a) elastic
modulus=200kPa and density= 60kg/m3 (b) elastic modulus=20kPa and
density=40kg/m3 for erect immobilized continuous sitting
Figure 4.6 shows the morphological changes at ischial tuberocity due to von
Mises stress in soft muscle tissues (for model II) for different continuous sitting time
interval. As depicted from the Figure 4.6, the size of damage area increases
continuously with time.
Figure 4.6 (a) Finite element model for soft seat cushion (E=200kPa and  =60kg/m3)
with magnification of muscle region subjected to high stress
64
Figure 4.6 (b) Stress distribution at ischial tuberocity for erect immobilized continuous
sitting on soft cushion (E=200kPa and  =60kg/m3) for 1800 sec (von
Mises stress=22002Pa High stress area=1.24 mm2)
Figure 4.6 (c) Stress distribution at ischial tuberocity for erect immobilized continuous
65
Figure 4.6 (d) Stress distribution at ischial tuberocity for erect immobilized continuous
Figure 4.7 shows a von Mises stress distribution in soft tissue muscles for a
cushion having elastic modulus of 20kPa and density of 40 kg/m3. Comparing Figures
4.6 and 4.7, it can be observed that the maximum stress affected area is less in case
of soft cushion of elastic modulus of 20kPa and density 40 kg/m 3 than that for the
cushion having elastic modulus of 200kPa and density 60kg/m3. Model II (E=20kPa
and density40kg/m3) predicts a maximum stress of 20324Pa at ischial tuberocity after
a continuous sitting duration of half an hour and the value approaches to the
maximum experimental value (nearly 19500Pa) (Verver et al., 2004). But the trend of
increasing stress at ischial tuberocity is nearly similar for both the cushion.
Comparing Figures 4.4, 4.6 and 4.7, it can be deduced that the size of the affected
zone as well as the stress is much larger for a rigid seat (Model I) as compared to
that of soft cushion (Model II). It is evident from Figures 4.6 and 4.7 that the cushion
having elastic modulus 20kPa and density 40kg/m3 shows less stress distribution at
ischial tuberocity in comparison to cushion having elastic modulus 200kPa and
density 60kg/m3. Therefore, further analysis carried out on soft cushion having elastic
modulus 20kPa and density 40kg/m3.
66
Figure 4.7 (a) Finite element model for soft seat cushion (E=20kPa and  =40kg/m3)
with magnification of muscle region subjected to high stress
Figure 4.7 (b) Stress distribution at Ischial tuberocity for erect immobilized continuous
Mises stress=20324Pa High stress area=0.17mm2)
67
Figure 4.7 (c) Stress distribution at Ischial tuberocity for erect immobilized continuous
Figure 4.7 (d) Stress distribution at Ischial tuberocity for erect immobilized continuous
The above discussions demonstrate dependence of von Mises stress at ischial
tuberocity on time of erect immobilized continuous sitting for different properties of
seats. Frequent immobilized continuous sitting makes the muscle tissues intolerable
for further compression even for soft polyurethane foam. Also in an erect immobilized
sitting, a transverse load is acted along the direction of fibers of muscles tissue
68
present below ischial tuberocity (Verver et al., 2005). Therefore, in order to make the
stress relax or to reduce the direct load on fibers either a postural changes or a
leisure time should be preferred after certain duration of sitting. Different posture
showing sidewise leaning are illustrated in Figure 4.8. Lifting from seat or a leisure
time after certain duration of continuous sitting make the body relaxes somewhat but
does not recover the muscle stiffness again back to starting zero level of relaxation.
During vertical sitting, the longissimus and gluteus muscles are loaded by the sacrum
and ischial tuberosities. Since the postural change causes change in position of the
ischial tuberocity, the longissimus and gluteus muscles are not subjected to that
much vertical load as in case of neutral position. Only a fraction of total vertical load
acts on ischial tuberocity and hence on the muscles just lies below it.
Load Load Load
b c
a
Figure 4.8 Postures (a) neutral position (b) left wise leaning (c) right-wise leaning
Postural change transfers the load to those regions of body other than on ischial
tuberocity and relieves the muscles below ischial tuberocity by distributing the stress
towards thighs and waist. The analysis for posture shown in Figure 4.8. It has been
observed that stress drops from 37193Pa to 34500Pa with a change in tilting angle
from 00 (erect sitting) to 300 while interacting with rigid seat for half an hour. The
stress drops from 20324Pa to 14632Pa for soft cushion of elastic modulus of 20kPa
with density 40kg/m3 for change in tilting angle from 00 (erect sitting) to 300. However,
sitting continuously with same tilting angle (300) or posture from thirty minutes to
three hours, the stress increases from 34500Pa to 35600Pa while interacting with
rigid seat. Similarly, sitting continuously with same tilting angle (300) or posture from
thirty minutes to three hours, the stress increases from 14632Pa to 16922Pa while
interacting with soft cushion of elastic modulus 20kPa and density 40kg/m 3. Stress
distribution is shown in Figure 4.9. Figure 4.10 displays a maximum shear stress of
8438Pa after thirty minutes of continuous sitting with 300 tilting posture and the shear
stress increases gradually when the time increases.
69
Figure 4.9 (a) Finite element model for immobilized continuous sitting with tilting
angle of 300 for soft cushion with elastic modulus of 20kPa and density
40kg/m3
Figure 4. 9 (c) Stress distributions at

Figure 4.9 (b) Stress distributions at
ischial tuberocity on immobilized
ischial tuberocity on immobilized
continuous sitting with tilting angle of 300
continuous sitting with tilting angle of 300
for soft cushion of elastic modulus of
for soft cushion of elastic modulus of
20kPa and density 40kg/m3 (von Mises
20kPa and density 40kg/m3 (von Mises
stress=15069Pa Time=3600sec)
stress=14632Pa Time=1800sec)
Figure 4.9 (e) Stress distributions at

Figure 4.9 (d) Stress distributions at ischial tuberocity on immobilized
ischial tuberocity on immobilized continuous sitting with tilting angle of 300
continuous sitting with tilting angle of 300 for soft cushion of elastic modulus of
for soft cushion of elastic modulus of 20kPa and density 40kg/m3 (von Mises
20kPa and density 40kg/m3 (von Mises stress=16922Pa Time=7200sec)
stress=15975Pa
Time=5400sec)
70
Figure 4.9 (f) Stress distributions at ischial Figure 4.9 (g) Stress distributions at
tuberocity on immobilized continuous ischial tuberocity on immobilized
sitting with tilting angle of 300 for soft continuous sitting with tilting angle of 300
cushion of elastic modulus of 20kPa and for soft cushion of elastic modulus of
density 40kg/m3 (von Mises 20kPa and density 40kg/m3 (von Mises
stress=16395Pa Time=9000sec) stress=16922Pa Time=10800sec)
Figure. 4.10 (a) Finite element model for immobilized continuous sitting with tilting
angle of 300 for soft cushion with elastic modulus of 20kPa and density
40kg/m3
Figure 4.10 (b) Shear stress distribution at Figure 4.10 (c) Shear stress distribution at
ischial tuberocity on immobilized ischial tuberocity on immobilized
continuous sitting with tilting angle of 300 continuous sitting with tilting angle of 300
for soft cushion of elastic modulus 20kPa for soft cushion of elastic modulus 20kPa
and density 40kg/m3 (Shear and density 40kg/m3 (Shear
71
Figure 4.10 (d) Shear stress distribution at Figure 4.10 (e) Shear stress distribution
ischial tuberocity on immobilized at ischial tuberocity on immobilized
stress=8738Pa stress=8922Pa
Time=5400sec) Time=7200sec)
Figure 4.10 (f) Shear stress distribution at Figure 4.10 (g) Shear stress distribution
ischial tuberocity on immobilized at ischial tuberocity on immobilized
The pattern of change of von Mises stress and shear stress at ischial tuberocity is
analyzed for change in cushion thickness as shown in Figures 4.11a and 4.11b
respectively. It has been found that increase of cushion thickness is effective in
reducing von Mises stress for any time duration of continuous sitting. However, shear
stress increases with thickness of cushion. The thickness of cushion is increased up
to eighty mill meters since the cushion thickness beyond this value is ineffective in
decreasing the stress beneath the ischial tuberocity (Chow and Odell., 1978).
72
Soft cushion (E=20kPa and density=40 kg/m3) at an angle of tilting 300
22000

20000
18000
16000
14000
12000
10000
1800 3600 5400 7200 9000 10800
(a)
Soft cushion (E=20kPa and density =40kg/m3 ) at an angle of tilting 300
9500
Shear stress in Pascal
9000
8500
8000
7500
7000
1800 3600 5400 7200 9000 10800
60 mm 70 mm 80 mm
(b)
Figure 4.11 Effect of time duration of tilted (300) sitting on (a) von Mises stress (b)
shear stress for soft cushion (E= 20kPa and density=40kg/m3)
4.4 Conclusions
The present study presents a numerical approach for analyzing the stresses
being developed beneath ichibial tuberocity due to continuous working in office
environment in sitting posture on office chair. The analysis provides insight into the
problem and suggests the ways to reduce the stress on bony prominence causing
73
cell death of muscle tissue. The methodology provides guidelines to avoid suffering
from pressure sore to some extent in an office environment. Although analysis shows
that stress within muscle decreases with proper cushion thickness and postural
changes but a postural change always deals with the change in position of ischial
tuberocity during leaning (Gefen et al., 2005). In order to overcome this limitation, the
present study deals with calculation of stress beneath the ischial tuberocity. It has
been shown that use of right kind of foam for seat cushion and thickness can
substantially reduce the stress level at ischial tuberocity. This work considers a
simple 2D formulation to provide guidelines for the designers to analyze behavior of
interaction of soft human tissue and cushion material. The study can be improved by
considering a real model with 3D formulation.
74
CHAPTER 5
A Novel Multi-Attribute
Decision Making
Approach for Product
Selection Conforming
Ergonomic
Considerations
5.1 Introduction
Competition in the market place demands high workload in an office environment
resulting in prolonged sitting. Prolonged sitting may cause health risk like muscular
disorders (Hales and Bernard, 1996). Therefore, ergonomically designed office chair
possessing capability of maintaining compatibility between the user and product may
lead to reduce fatigue. While designing or procuring an office chair, the psychological
needs must be fulfilled in addition to physical needs to improve user satisfaction.
Selection of an office chair with salient features satisfying ergonomic needs (both
physical and psychological needs) becomes a complex decision making process.
Keeping in view of complexity of the problem, multi-attribute decision making (MADM)
approach can be considered during product design focusing on the requirements of
user in terms of conflicting criteria in order to solve the task of selection of an
ergonomically designed product. In a decision making process, it is unlikely that
decision makers can express their preferences using crisp rating for attributes (Jee
and Kang, 2000; Shanian and Savadogo, 2006; Jahan et al., 2010). As experts are
not able to exactly specify to their preferences, linguistic variables using a fuzzy scale
is used to conveniently deal with impreciseness and ambiguity in judgement (Chen,
2000; Chen et al., 2006; Girubha and Vinodh, 2012 ). Decision making in fuzzy
environment has been suitably articulated by Zadeh (1965), Zadeh and Bellman
(1970) and Carlsson and Fuller (1996). Still the decision making becomes
inconsistent because most of the approaches consider either objective or subjective
attributes (Rao, 2012; Maniya and Bhatt, 2010). In decision making, usually some
attributes are objective and some are subjective in nature. The attributes need to be
properly evaluated for estimating attribute weights integrating both objective and
subjective criteria (Rao and Patel, 2010). The subjective attributes can be dealt using
eigen method (Saaty, 1977) or Delphi method (Hwang and Lin, 1987) whereas the
objective attributes can be effectively managed by entropy method (Hwang and
Yoon, 1981) for weight estimation. To address this issue, a novel decision making
technique is proposed in this work considering both subjective and objective weights
for attributes in order to facilitate the decision maker to deal with objective information
regarding the product as well as the uncertainty of human judgment. The attribute
ratings obtained from multiple experts are aggregated for effective decision making.
Three different popular MADM methods such as TOPSIS (Techniques for Order
Preference by Similarity to Identical Solution), VIKOR (VIseKriterijumska Optimizacija
I Kompromisno Resenjea) and PROMETHEE (Preference Ranking Organization
Method for Enrichment Evaluations) are used to solve the selection problem of
choosing the best ergonomically designed office chair. All the methods are
75
considered under same managerial decision. TOPSIS, a linear weighting technique
developed by Hwang and Yoon (1981) in its crisp form and then in expanded form by
Chen and Hwang (1992), is based on the concept that the chosen alternative should
have the shortest Euclidean distance from positive ideal solution and farthest from
negative ideal solution (Rao, 2006). VIKOR, on the other hand, determines a
compromise solution which is feasible and closest to the ideal solution but makes an
agreement by mutual concession which the help of the decision maker to take a
decision with conflicting criteria (Opricovic and Tzeng, 2004). It introduces the multi-
criteria ranking index based on the particular measure of ‘‘closeness” to the ‘‘ideal”
solution (Opricovic, 1998). PROMETHEE proceeds to a pair-wise comparison of
alternatives in each single criterion in order to determine partial binary relations
denoting the strength of preference of an alternative ‘a’ over alternative ‘b’ (Rao and
Rajesh, 2009). In order to check the stability of ranking with respect to different
weighted attributes, a sensitivity analysis has been performed considering different
proportion of attribute weight (subjective and objective) and the evaluation carried
under three different MADM methods.
5.2 Proposed Methodology

MADM has established as an effective methodology for solving a large variety of
multi-criteria decision making and ranking problems (Hwang and Yoon, 1981). In this
study, a novel approach of MADM has been proposed to find a suitable
ergonomically designed product with respect to design characteristics (attributes).
The best alternative is chosen from a set of n alternatives {A1,A2, …, An} whereas the
performance of the alternatives are decided on the basis of m attributes {C1, C2, …,
Cm} by a group of k decision maker (DMs) {DM1, DM2, …, DMk}as given in Table 5.1.
The weight for the attributes are considered as {w1, w2, ….wm}.
Table 5.1 Decision matrix
Alternatives Attributes
C1 C2 - - Cm
(w1) (w2) (wm)
A1 *
x11 *
x12 - - *
x1m
A2 *
x 21 *
x 22 - - *
x 2m
- - - - - -
- - - - - -
An x n*1 x n* 2 - - *
x nm
Different steps of the proposed work are described in Figure 5.1. The
methodology consists of six major computational steps as discussed below.
76
Figure 5.1 Graphical representation of generic MADM model
Step 1: Identify the problem

To illustrate the approach, an ergonomically designed office chair selection
problem is considered as a case study. As the selection of an ergonomically
designed office chair includes technical specification as well as user preference, it
becomes a difficult task to choose an office chair with specific features that satisfy a
range of customers. In addition to design characteristics in terms of technical
specifications provided by manufacturers, expert opinion is also considered to take
into account the customer preferences. A group of DMs analysed the possible
attributes and alternatives from a set of available office chairs in the market place and
important attributes and alternatives are considered. Since impreciseness and
ambiguity exist to assign rating for each attribute and alternative, a linguistic scale is
used to express decision makers’ opinion on each alternative with respect to
attribute.
Step 2: Preparation of decision table
To model decision makers’ judgement, fuzzy scales are employed which translate
the linguistic terms into triangular fuzzy numbers as linguistic variables deal with
ambiguity and subjectivity (Zadeh, 1975). To convert the qualitative terms into
quantitative values, a five point fuzzy scale with triangular fuzzy numbers based on
the works of Chen (1985) is chosen. As shown in Figure 5.2, linguistic terms “very
77
low”(VL), “low”(L), “medium”(M), “high”(H), “very high”(VH) are included to measure
the performance of each alternative with respect to each attribute. The crisp score of
fuzzy number ‘M’ is obtained as follows (Chen, 1985):
x , 0  x  1
 max ( x )   (5.1)
0, otherwise
1  x, 0  x  1
 min ( x )   (5.2)
0, otherwise
The fuzzy max and fuzzy min of fuzzy numbers are defined in a manner such that
absolute location of fuzzy numbers can be automatically incorporated in the
comparison case. The left score of each fuzzy number ‘Mi’ is defined as
μ L M i   Supμ min x   μ Mi x  (5.3)
x
The μ L M i  score is a unique, crisp, real number in (0, 1). It is the maximum
membership value of the intersection of fuzzy number Mi and the fuzzy min. The right
score is obtained as:
μ R M i   Supμ max x   μ Mi x  (5.4)
x
Again μ R M i  is a crisp number (0, 1). Given the left and right scores, the total crisp
score of a fuzzy number Mi is defined as:
 T M i    R M i   1   L M i  / 2 (5.5)
These ratings may be given by a single or a group of decision maker. Yue (2011)
states that MADM problems can provide reliable results if analysis of multiple experts
is taken into account instead of the analysis of a single expert.
VL L M H VH
1
0.9
0.8
Membership function
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.25 0.5 0.75 1
0 Fuzzy number
Figure 5.2 Linguistic terms to fuzzy numbers conversion (5-point scale)
78
After evaluation of aggregate crisp score value (xij), the rating value for alternative
i with respect to each attribute j, the aggregate crisp values are normalized so that
rating values given by the decision makers can be converted into a common scale.
Considering the normalized value ( x *ij ), a decision table is prepared. Normalization is
carried out using following relationship.

x ij
x *ij 
x  ij max
(5.6)
where, xij is the aggregate crisp score of alternative ‘i’ under attribute ‘j’.
Step 3: Allocate the weights of importance of the identified attributes
The proposed methodology uses integrated weights of objective and subjective
preference for assigning attribute weights. By varying proportion of objective and
subjective weights, a large number of decision making scenarios can be generated to
provide the decision makers a wide range of solutions to choose the best one.
Step 3(a): Computation of objective weights of importance of the attributes

The objective weights can be computed by using the normalized data given in
decision matrix developed in previous step. As the statistical variance gives a
measure of dispersion of data points around their mean value (Rao and Patel, 2010),
the proposed method determines the objective weight of attributes in terms of
statistical variance method. The statistical variance for determining the objective
weights of importance of the attributes is given by the following equation.
2
   
n
v j  1 / n  x *ij  x *ij mean (5.7)
i 1
where vj is the variance of the data corresponding to the jth attribute and x *ij  
mean
is the
average value of x *ij .
The objective weight of the jth attribute, w 0j can be computed by dividing the
statistical variance of the jth attribute with the total value of the statistical variances for
m number of attributes. Thus, w 0j can be computed by the following equation.
vj
w 0j  m (5.8)
v
j1
j
Step 3(b): Computation of subjective weights of importance of the attributes

The subjective preferences can be evaluated through pair-wise comparison of the
attributes. A pair-wise comparison matrix (m × m) for all attributes can be constructed
with respect to objective by using Saaty's 1-9 scale of pair-wise comparisons so that
79
each attribute can compared with each other attribute. For each comparison, the
decision maker decides which of the attribute is most important among two and then
assigns a score to show how much more important it is than the other. After making
the pair-wise comparisons, the consistency is checked by using the following
computations.
 max  m
Consistency Index, CI  (5.9)
m 1
where, λmax is the maximum Eigen value of the matrix and m is matrix size
CI
Consistency Ratio, CR  (5.10)
RI
Consistency ratio (C.R.) can be defined as the ratio of consistency index (C.I.) and
randomly generated consistency index (R.I.) values (Saaty, 1980). The judgement
matrix is consistent if a CR value is less than 0.10.
Step 4: Computation of integrated weights of importance of the attributes

For utilizing both objective and subjective weights of the attributes, an integrated
weight of importance is to be calculated. The integrated weight can be described by
using the following equation.
w ij  w o  w oj  w s  w sj (5.11)
where, w ij , w oj and w sj denote the integrated, objective and subjective weight of the jth
attribute respectively. The weightings are taken in between 0 and 1. w o and w s

represent the weightings proportion considered for objective and subjective weights
respectively. The weightings are taken between 0 and 1.
Step 5: Determination of ranking of the alternatives

Each decision matrix has three main components viz., (a) alternatives, (b)
attribute, (c) weight or relative importance of each attribute. Three different MADM
methods such as Techniques for Order Preference by Similarity to Identical Solution
(TOPSIS), a compromise ranking method known as VIseKriterijumska Optimizacija I
Kompromisno Resenjea (VIKOR) and Preference Ranking Organization Method for
Enrichment Evaluations (PROMETHEE) are considered to measure the performance
of alternative. The selected MADM methods adopt different strategies for ranking the
alternatives. TOPSIS ranks the alternative based on shortest distance from the
positive ideal solution and the farthest distance from the negative-ideal solution
whereas VIKOR method provides a compromising solution in which an agreement is
established between two mutual concessions. PROMETHEE method uses ‘‘net
80
preference flow’’ function to rank the alternatives (Brans et al., 1984). Both VIKOR
and PROMETHEE methods use linear normalization as shown in equation 5.6.
TOPSIS method uses vector normalization and the normalized value can be obtained
by following equation.
1
x  x ij
*  m 2
2
ij  j1 x ij 
(5.12)
5.2.1 Techniques for Order Preference by Similarity to Identical Solution
(TOPSIS)
TOPSIS method is based on calculation of preference index in order to evaluate
the ranking of alternatives by computing the shortest Euclidean distances to both
positive ideal solution and negative ideal solution simultaneously. It is based on the
idea that the chosen alternative should have the shortest distance from the positive
ideal solution and on the other hand, the farthest distance from the negative ideal
solution. Here, the normalized decision matrix can be obtained by equation 5.13.
1
x *ij  x ij   x ij 
m 2
2
(5.13)
 j1 
The weighted normalized decision matrix considering integrated weights can be
expressed as Yij  W ij  X *ij . The positive ideal solution and negative ideal solution can
be calculated using following formulae.

The positive ideal (best) solutions can be expressed as:

A* = y1* ,..........., y *m 
 '  '' 
 max y ij j  J ,  min y ij j  J 
 i  i  (5.14)
The negative ideal (worst) solutions can be expressed as:

A - = y1- ,..........., y m 
 '  '' 
 min y ij j  J ,  max y ij j  J 
 i  i  (5.15)
where J ' is associated with beneficial attribute and J " is associated with non-
beneficial attribute.
The separation of each alternative from the ideal one is given by the Euclidean
distance. The separation of each alternative from the positive ideal solution is given
as
 y  , i 1,....I
m
2
D *i  ij  y *j (5.16)
j1
81
Similarly, the separation of each alternative from the negative ideal solution is
expressed as
 y  , i 1,....I
m
2
D i  ij  y j (5.17)
j1
After calculating the separation the relative closeness to the ideal solution is
carried out. The relative closeness (preference index) of the alternatives to the ideal
solution is defined as
D i
C*i  , i  1,....I (5.18)
D *i  D i
C *i is also called as the overall performance score of alternative. A set of alternatives
is generated according to the value of C *i indicating the most preferred and least
preferred feasible solutions. The alternative which has highest value of performance
score will be given top ranking in the order. Ranking will be done for different
proportion of subjective and objective weights. The final selection of the best
alternative will be assessed through the analysis of final ranking matrix.
5.2.2 VIseKriterijumska Optimizacija I Kompromisno Resenjea (VIKOR)

The foundation for compromise solution was established by Yu (1973) and Zeleny
(1982) and later advocated by Oprocovic and Tzeng (2002, 2007) and Tzeng et al.
(2002, 2005). The compromise solution is closest to the ideal solution which is a
feasible solution. The compromise ranking algorithm of the VIKOR method has the
following steps:
For alternative Ai the rating of jth attribute is expressed as fij
*
Step 1: The first step is to determine the objective, also determine the best, i.e., f j

and the worst, i.e. f j , values of all attributes.
f j*  max f ij ,j=1,2,……m (5.19(a))

i
f j  min f ij ,j=1,2,…….m (5.19(b))

i
Step 2: Compute the values Sj and Rj, i=1,2,…,n.
Si   w j
n f *
 f ij 
f 
j
(5.20)
j1
*
j  f j
R i  max w j
f *
 f ij 
f 
j
(5.21)
j
*
j  f j
where wj, are the weights of the attribute expressing the relative importance.
82
Step 3: Compute the values Qi, i=1,2,…n by the following relation

Qi   Si  S*  S 
 
 S*  1   R i  R *  R 
 R*  (5.22)
where S* is the minimum value of S i i.e. S  min Si and S  is the maximum value of
*
i

Si i.e. S  max Si
i
Similarly, R * is the minimum value of the R i i.e. R  min R i and R  is the

*
i

maximum value of Ri i.e R  max R i
i
 is introduced as the weight of strategy of “ the majority of attribute” (or the

maximum group utility”), usually   0.5 .
Step 4: By arranging the alternatives in the ascending order of S, R and Q values,
the three ranking lists can be obtained. The compromise ranking list for a given  is
obtained by ranking with Qi measures. The best alternative, ranked by Qi, is the one
with the minimum value of Qi.
Step 5: Propose a compromise solution for alternative Ak
Under a given weight of attribute, alternative Ak is the best ranked by Q value
(Minimum) if the following two conditions are satisfied (Tzeng et al., 2005):
Condition 1: ‘Acceptable advantage’: QA K   QA1   DQ (5.23)
DQ  1 N  1 (5.24)
where, A 1 the second best alternative in the ranking list by Q. N is the number of
alternatives.
Condition 2:‘Acceptable stability in decision making’: Alternative Ak must also be the
best ranked by S or/and R. This compromise solution is stable within a decision
making process, which could be ‘‘voting by majority rule’’ (when   0.5 is needed),
or ‘‘by consensus’’ (   0.5 ), or ‘‘with veto’’   0.5 . Here,  is the weight of the
decision making strategy ‘‘the majority of attribute’’ (or ‘‘the maximum group utility’’).
If one of the conditions is not satisfied, then a set of compromise solutions is
proposed, which consists of:
1- Alternatives A K and A 1 if only condition 2 is not satisfied
2- Alternatives A K , A 1 ,….. A P if condition 1 is not satisfied; A P is determined by the
relation QA P   QA1   DQ
83
5.2.3 Preference Ranking Organization Method for Enrichment Evaluations
(PROMETHEE)
PROMETHEE method was introduced by Brans et al. (1984). Like all other
ranking methods, PROMETHEE deals with a pair wise comparison of alternatives for
each single attribute in order to determine partial binary relations denoting the
strength of preference of an alternative A1 over alternative A2. The alternatives are
evaluated on different attribute. The implementation of PROMETHEE also requires
relative importance or the weights of the attribute considered and information on the
decision maker preference function, which he/she uses when comparing the
contribution of the alternatives in terms of each separate attribute. The preference
function (Pi) translates the difference between the evaluations obtained by two
alternatives (A1 and A2) in terms of a particular attribute, into a preference degree
ranging from 0 to 1.The method covers the following steps given below
Here the decision maker gives his/her preference function by comparing the
contribution of one alternative with respect to another in terms of each separate
attribute. The preference function (Pi) finds a difference between two alternatives (A1
andA2) for a particular attribute in terms of a preference degree 0 or 1. Let P j, A1A2 be
the preference function associated to the attribute Cj.
 
Pj,A1A 2  G j C j A1   C j A 2  (5.25)
0  Pi,A1A 2  1
where, Gj is a non-decreasing function of the observed deviation (d) between two
alternativesA1 and A2 over the attribute Cj. Let the decision maker have specified a
preference function Pj and weight wj for each attribute Cj(j=1, 2, . . . , m). The multiple
attribute preference index  A1A 2 is then defined as the weighted average of the
preference functions Pj
m
 A1A 2   w j Pj.A1A 2 (5.26)
j1
 A1A 2 , represents the intensity of preference of the decision maker of alternative

A1over alternative A2 when considering simultaneously all the attribute and the value
ranges from 0 to 1. This preference index determines a valued outranking relation on
the set of actions. As an example, the schematic calculation of the preference indices
for a problem consisting of three alternatives and four attribute is given in Figure 5.3
(Marinoni, 2005).
84
Figure 5.3 Preference indices for a problem consisting of three alternatives and four
attributes.
For PROMETHEE outranking relations, the leaving flow, entering flow and the net
flow for a particular alternative(A) belonging to a set of alternatives ‘A’ are defined by
the following equations:
  (A)    xa (5.27)
xA
  (A)    ax (5.28)
xA
(A)    (A)    (A) (5.29)
  (A) is called the leaving flow,   (A) is called the entering flow and (A) is called
the net flow.   (A) is the measure of the outranking character of A (i.e. dominance of
alternative ‘A’ overall other alternatives) and   (A) gives the outranked character of
A(i.e. degree to which alternative A is dominated by all other alternatives). The net
flow, (A) represents a value function, whereby a higher value reflects a higher
attractiveness of alternative A. The net flow values are used to indicate the
outranking relationship between the alternatives. For example, for each alternative A,
belonging to the set of alternatives (A1, A2….An),  A1A 2 is an overall preference index
of A1 over A2, taking into account all the attribute. Alternative A1 outranks A2 if
A1   A 2  and A1 is said to be indifferent to A2 if A1   A 2  .
85
Keeping view with the increasing demand for a suitable ergonomically designed
office chair, six different alternatives with respect to ten design characteristics
(attribute) are considered as shown in Figure 5.4. The attribute are considered with
an extensive literature review of previous report for ergonomically designed office
chair (Mohanty and Mahapatra, 2014). A survey among manufacturers of chairs and
opinion of experts specialized in ergonomically designed chair revealed that the
evaluation of office chair should carry ten important design characteristics. Based on
the evaluation of four decision makers, a decision matrix is made considering six
alternatives {A1, A2, A3, A4, A5, A6} and ten attributes {C1, C2… C10}. Relative weights
{w1, w2 ….w10} are assigned to each design characteristic (attribute) to represent the
DM’s preference information.
Figure 5.4 Office chair model for analysis
Ten important attributes considered are Depth of seat (C1), Overall depth (C2),
Width of seat (C3), Size of base (C4), Width height ratio (C5), Seat adjustment (C6),
Backrest height (C7), Swivel angle (C8), Decoration (C9), and Density of cushion (C10).
Based on the dimensions considered and comparison with Bureau of Indian standard
data, the attributes are classified into beneficial and non-beneficial category. Out of
ten attribute, C1, C2, C3, C4, C5, C6, C7, C8 and C9 are beneficial (higher the value is
86
desired) and C10 is non-beneficial attribute (lower the value is desired). As the
alternatives (Office chair) based on attributes are of conflicting in nature, a five point
fuzzy scale with triangular fuzzy numbers is chosen to rate the alternatives. A team of
four decision makers, DM1, DM2, DM3 and DM4 has been formed to evaluate the
alternatives. An individual decision maker’s judgment is evaluated by using fuzzy
rating scale with triangular membership functions in order to extract the rating values
of alternatives where individual attribute is given linguistic terms as is given in Table
5.2. Linguistic terms are further converted to their corresponding fuzzy numbers as
shown in Table 5.3. In order to assess with each attribute weight, individual fuzzy
numbers are aggregated as is highlighted in Table 5.4. Aggregate fuzzy numbers are
then transformed into crisp values and the corresponding values are given in Table
5.5.
Table 5.2 Linguistic rating for alternatives selection
Decision Alternative C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Maker
D1 A1 L L VL M L H M M VL L
A2 M M H H H M M M H L
A3 L H M M L L H VL VL M
A4 L VL L H M M H VL M M
A5 M H M M M M M L VH L
A6 L L VL M L H M L M L
D2 A1 M M L H M VH H H L M
A2 H M H H VH H H H M L
A3 M H H H M M VH L L H
A4 M M M VH M H H L M H
A5 VH H H M H H H M H L
A6 M M L M M VH M M M M
D3 A1 H M M M M VH H VH L L
A2 VH H VH VH VH VH VH VH H M
A3 H VH H M M H VH L VL H
A4 H M M VH H VH VH L M H
A5 VH VH H M H VH H H VH M
A6 H M M L M VH H M H L
D4 A1 VL L VL H L M M M M L
A2 M L M VH M M M L VH L
A3 L M M M L L H VL L M
A4 L VL L H L M H VL M M
A5 M M M M M M M L H L
A6 VL L VL L L M L VL M L
87
Table 5.3 Fuzzy numbers associated with alternatives
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
D1 A1 L(0,0.3,0.5) L(0,0.3,0.5) VL(0,0,0.3) M(0.3,0.5,0.7) L(0,0.3,0.5) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VL(0,0,0.3) L(0,0.3,0.5)
A2 M(0.3,0.5,0.7) M(0.3,0.5,0.7) H(0.5,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) H(0.5,0.7,1) L(0,0.3,0.5)
A3 L(0,0.3,0.5) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) L(0,0.3,0.5) H(0.5,0.7,1) VL(0,0,0.3) VL(0,0,0.3) M(0.3,0.5,0.7)
A4 L(0,0.3,0.5) VL(0,0,0.3) L(0,0.3,0.5) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) H(0.5,0.7,1) VL(0,0,0.3 M(0.3,0.5,0.7) M(0.3,0.5,0.7)
A5 M(0.3,0.5,0.7) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) VH(0.7,0.7,1) L(0,0.3,0.5)
A6 L(0,0.3,0.5) L(0,0.3,0.5) VL(0,0,0.3) M(0.3,0.5,0.7) L(0,0.3,0.5) H(0.5,0.7,1) M(0.3,0.5,0.7) L(0,0.3,0.5) M(0.3,0.5,0.7) L(0,0.3,0.5)
D2 A1 M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) H(0.5,0.7,1) M(0.3,0.5,0.7) VH(0.7,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) L(0,0.3,0.5) M(0.3,0.5,0.7)
A2 H(0.5,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) H(0.5,0.7,1) VH(0.7,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) L(0,0.3,0.5)
A3 M(0.3,0.5,0.7) H(0.5,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VH(0.7,0.7,1) L(0,0.3,0.5) L(0,0.3,0.5) H(0.5,0.7,1)
A4 M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VH(0.7,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) H(0.5,0.7,1) L(0,0.3,0.5) M(0.3,0.5,0.7) H(0.5,0.7,1)
A5 VH(0.7,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) L(0,0.3,0.5)
A6 M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VH(0.7,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7)
D3 A1 H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VH(0.7,0.7,1) H(0.5,0.7,1) VH(0.7,0.7,1) L(0,0.3,0.5) L(0,0.3,0.5)
A2 VH(0.7,0.7,1) H(0.5,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7)
A3 H(0.5,0.7,1) VH(0.7,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) H(0.5,0.7,1) VH(0.7,0.7,1) L(0,0.3,0.5) VL(0,0,0.3) H(0.5,0.7,1)
A4 H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) VH(0.7,0.7,1) H(0.5,0.7,1) VH(0.7,0.7,1) VH(0.7,0.7,1) L(0,0.3,0.5) M(0.3,0.5,0.7) H(0.5,0.7,1)
A5 VH(0.7,0.7,1) VH(0.7,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) VH(0.7,0.7,1) H(0.5,0.7,1) H(0.5,0.7,1) VH(0.7,0.7,1) M(0.3,0.5,0.7)
A6 H(0.5,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) M(0.3,0.5,0.7) VH(0.7,0.7,1) H(0.5,0.7,1) M(0.3,0.5,0.7) H(0.5,0.7,1) L(0,0.3,0.5)
D4 A1 VL(0,0,0.3) L(0,0.3,0.5) VL(0,0,0.3) H(0.5,0.7,1) L(0,0.3,0.5) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5)
A2 M(0.3,0.5,0.7) L(0,0.3,0.5) M(0.3,0.5,0.7) VH(0.7,0.7,1) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) VH(0.7,0.7,1) L(0,0.3,0.5)
A3 L(0,0.3,0.5) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) L(0,0.3,0.5) H(0.5,0.7,1) VL(0,0,.3) L(0,0.3,0.5) M(0.3,0.5,0.7)
A4 L(0,0.3,0.5) VL(0,0,0.3) L(0,0.3,0.5) H(0.5,0.7,1) L(0,0.3,0.5) M(0.3,0.5,0.7) H(0.5,0.7,1) VL(0,0,0.3) M(0.3,0.5,0.7) M(0.3,0.5,0.7)
A5 M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) M(0.3,0.5,0.7) L(0,0.3,0.5) H(0.5,0.7,1) L(0,0.3,0.5)
A6 VL(0,0,0.3) L(0,0.3,0.5) VL(0,0,0.3) L(0,0.3,0.5) L(0,0.3,0.5) M(0.3,0.5,0.7) L(0,0.3,0.5) VL(0,0,0.3) M(0.3,0.5,0.7) L(0,0.3,0.5)
88
Table 5.4 Aggregate fuzzy number of alternatives
Alternatives Criteria
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
A1 (0.2,0.375,0.625) (0.15,0.4,0.6) (0.075,0.2,0.45) (0.4,0.6,0.85) (0.15,0.4,0.6) (0.55,0.65,0.925) (0.4,0.6,0.85) (0.45,0.6,0.85 (0.075,.275,0.5) (0.075,0.35,0.55)
A2
(0.45,0.6,0.85) (0.275,0.5,0.725) (0.5,0.65,0.925) (0.6,0.7,1) (0.55,0.65,0.925) (0.45,0.6,0.85) (0.45,0.6,0.85) (0.375,0.55,0.8) (.5,0.65,0.925) (0.075,0.35,0.55)
A3
(0.2,0.45,0.675) (0.5,0.65,0.925) (0.4,0.6,0.85) (0.35,0.55,0.775) (0.15,0.4,0.6) (0.2,0.45,0.675) (0.6,0.7,1) (0,0.15,0.4) (0,0.15,0.4) (0.4,0.6,0.85)
A4
(0.2,0.45,0.675) (0.15,0.25,0.5) (0.15,0.325,0.55) (0.6,0.7,1) (0.275,0.5,0.725) (0.45,0.6,0.85) (0.55,0.7,1) (0,0.15,0.4) (0.3,0.5,0.7) (0.4,0.6,0.85)
A5
(0.5,0.6,0.85) (0.5,0.65,0.925) (0.4,0.6,0.85) (0.3,0.5,0.7) (0.4,0.6,0.85) (0.45,0.6,0.85) (0.4,0.6,0.85) (0.2,0.45,0.675) (0.6,0.7,1) (0.075,0.35,0.55)
A6 (0.2,0.375,0.625) (0.15,0.4,0.6) (0.075,0.2,0.45) (0.15,0.4,0.6) (0.15,0.4,0.6) (0.55,0.65,0.925) (0.275,0.5,0.725) (0.15,0.325,0.55) (0.35,0.55,0.775) (0.075,0.35,0.55)
Table 5.5 Crisp ratings of alternatives

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
A1 0.409574 0.41 0.268889 0.59 0.41 0.6582 0.59 0.600867 0.318665 0.36642
A2 0.60087 0.5 0.645354 0.702797 0.6582 0.60087 0.60087 0.554043 0.645354 0.36642
A3 0.45551 0.645354 0.59 0.545493 0.41 0.45551 0.702797 0.225217 0.225217 0.59
A4 0.45551 0.313636 0.362788 0.702797 0.5 0.60087 0.688963 0.225217 0.5 0.59
A5 0.612727 0.645354 0.59 0.5 0.59 0.60087 0.59 0.45551 0.702797 0.36642
A6 0.409574 0.41 0.268889 0.41 0.41 0.6582 0.5 0.362788 0.545493 0.36642
Table 5.6 Normalized crisp ratings

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
A1 0.668445 0.63531 0.416653 0.839503 0.622911 1 0.839503 1 0.453424 0.621051
A2 0.980648 0.774769 1 1 1 0.912898 0.854969 0.922072 0.918265 0.621051
A3 0.743415 1 0.914227 0.776175 0.622911 0.692054 1 0.374821 0.320459 1
A4 0.743415 0.485991 0.562153 1 0.759648 0.912898 0.980316 0.374821 0.711443 1
A5 1 1 0.914227 0.711443 0.896384 0.912898 0.839503 0.758088 1 0.621051
A6 0.668445 0.63531 0.416653 0.583383 0.622911 0.999999 0.711443 0.603774 0.776175 0.621051
89
The aggregate crisp values of attribute are now normalized using equation 5.6 so
that the attribute ratings given by the decision makers can be converted into a
common scale. The normalized decision matrix for attribute is shown in Table 5.6.
On the basis of statistical variance method, the variance and the objective
weights of the attributes are computed by using equations 5.7 and 5.8. The variance
and the objective weight value for ten attribute are given in Table 5.7.
Table 5.7 Objective weights of attribute
Attribute C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Variance 0.0189 0.0369 0.0602 0.0224 0.0220 0.0105 0.0093 0.0598 0.0580 0.0320
Objective weights
0.057 0.112 0.182 0.068 0.0668 0.032 0.028 0.181 0.176 0.096
Analytic Hierarchy Process (AHP) is used to calculate subjective weights of

attribute. A pair-wise comparison matrix ( 10 10 ) as shown in Table 5.8 can be
constructed for attribute based upon the intensity of importance. The value of CR is
calculated by using equations 5.9 and 5.10. The value of CR obtained is 0.0654
which is less than 0.1 and hence the result is acceptable. The subjective weights are
calculated using geometric means and the result is shown in Table 5.9.
Table 5.8 Pair wise comparison matrix.
Dept Overal Width of Size of Width Seat Back rest Swivel Decoration Densit
h l depth seat base height adjustmen height angle y
ratio t
Depth 1 3 2 3 1 2 4 3 6 1/4
Overall depth 1/3 1 1/3 1/4 1/6 1/5 1/2 1/5 2 1/5
Width of seat ½ 3 1 3 1/3 2 4 3 5 1/4
Size of base 1/3 4 1/3 1 1/3 1 4 1/3 4 1/2
Width height 1 6 3 3 1 4 5 3 6 1/4
ratio
Seat adjustment ½ 5 1/2 1 1/4 1 2 1/3 3 1/5
Backrest height ¼ 2 1/4 1/4 1/5 1/2 1 1/4 1/2 1/6
Swivel angle 1/3 5 1/3 3 1/3 3 4 1 3 1/3
Decoration 1/6 1/2 1/5 1/4 1/6 1/3 2 1/3 1 1/6
Density 4 5 4 2 4 5 6 3 6 1
To check the consistency of matrix eigen value λmax is to be calculated

Consistency index (CI):
 max  n 10.86  10

n 1 10  1 = 0.0956
Consistency ratio (CR):
CI 0.0956
= RI = 1.45 = 0.0654 < 0.1
90
Table 5.9 Subjective weight design attributes
Criterion Weight
1 Depth seat pan 0.143
2 Overall depth 0.027
3 Width of seat 0.110
4 Size of base 0.065
5 Width height ratio 0.188
6 Seat adjustment 0.0611
7 Backrest height 0.0268
8 Swivel angle 0.0881
9 Decoration 0.0259
10 Density 0.261
The integrated weights of attributes are obtained using equation 5.11. Table 5.10
gives the integrated weights of attributes considering the different weightings
proportion of the objective and subjective weights within the range 0 to 1.
91
Table 5.10 Integrated Weight calculation
Integrated Weights of Attribute

Importance of Importance of Attributes
Objective weight Subjective weight C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
(w0) (ws)
w ic1 w ic 2 w ic3 w ic 4 w ic5 w i
c6 w i
c7 w i
c8 w i
c9 w ic1 0
1.0 0 0.057 0.111 0.182 0.068 0.066 0.032 0.028 0.181 0.175 0.096
0.8 0.2 0.074 0.094 0.167 0.067 0.091 0.037 0.028 0.162 0.145 0.129
0.6 0.4 0.091 0.077 0.153 0.066 0.115 0.043 0.027 0.143 0.115 0.162
0.5 0.5 0.100 0.069 0.146 0.066 0.127 0.046 0.027 0.134 0.100 0.178
0.4 0.6 0.108 0.060 0.138 0.066 0.139 0.049 0.027 0.125 0.085 0.195
0.2 0.8 0.125 0.043 0.124 0.065 0.163 0.055 0.027 0.106 0.055 0.228
0 1.0 0.143 0.027 0.110 0.065 0.188 0.061 0.026 0.088 0.025 0.261
The normalized decision matrix for TOPSIS is obtained using equation 5.13 and
the decision matrix is shown in Table 5.11. The ranking of the alternatives is
illustrated by considering purely subjective weight (wo = 1 and ws = 0). By multiplying
normalized decision matrix with corresponding integrated attribute weights, the
weighted normalized decision matrix can be obtained as is given in Table 5.12.
Table 5.11 Normalized decision matrix
Alternatives Attribute
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
A1 0.336 0.333 0.228 0.411 0.330 0.448 0.391 0.570 0.251 0.330
A2 0.493 0.406 0.548 0.490 0.531 0.410 0.398 0.526 0.508 0.330
A3 0.373 0.524 0.501 0.381 0.330 0.310 0.466 0.214 0.177 0.531
A4 0.373 0.255 0.308 0.490 0.403 0.410 0.456 0.214 0.394 0.531
A5 0.502 0.524 0.501 0.350 0.476 0.410 0.391 0.432 0.553 0.330
A6 0.335 0.333 0.228 0.286 0.330 0.449 0.331 0.344 0.430 0.330
The positive ideal solution and the negative ideal solution for the alternatives are
calculated using equations 5.14 and 5.15 respectively. The positive ideal solution is
given as {0.0288, 0.058, 0.099, 0.033, 0.035, 0.014, 0.013, 0.103, 0.097, and 0.032}.
Similarly, the negative ideal solution is given by {0.019, 0.028, 0.041, 0.019, 0.022,
0.009, 0.009, 0.038, 0.031, and 0.051}. The positive and negative separation ( D *i and
D i ) of each alternative from ideal solutions is calculated using equations 5.16 and
5.17 respectively. The Preference index ( C *i ) showing the ranking of alternatives can
be obtained by using equation 5.18. and the final ranking of six alternatives are
depicted in Table 5.13. In the similar manner, the ranking order of six alternatives
considering integrated weights of different proportions is given in Table 5.14.
92
Table 5.12 Weighted normalized matrix for alternatives
Alternatives Attributes
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
A1 0.019 0.037 0.042 0.028 0.022 0.014 0.011 0.103 0.044 0.032
A2
0.028 0.045 0.099 0.033 0.035 0.013 0.011 0.095 0.090 0.032
A3
0.021 0.058 0.091 0.026 0.022 0.009 0.013 0.038 0.031 0.051
A4
0.021 0.028 0.056 0.033 0.027 0.013 0.013 0.038 0.070 0.051
A5 0.029 0.058 0.091 0.0240 0.031 0.013 0.011 0.078 0.097 0.032
A6 0.020 0.037 0.041 0.020 0.022 0.014 0.009 0.062 0.075 0.032
Table 5.13 Ranking index ( C *i ) of alternatives
Alternatives Positive Negative Preference Ranking of the

Separation Separation Index alternatives
Measure Measure
D *i D i C *i
A1 0.0057 0.0043 0.4389 3
A2 0.0002 0.0093 0.9762 1
A3 0.0076 0.0028 0.2692 5
A4 0.0070 0.0014 0.1692 6
A5 0.0006 0.0080 0.9213 2
A6 0.0055 0.0025 0.3153 4
Table 5.14 Ranking of alternatives considering integrated weight

Alternatives wo = 1 wo = 0.8 wo = 0.6 wo = 0.4 ws wo = 0.2 wo = 0
ws = 0 ws = 0.2 ws = 0.4 = 0.6 ws = 0.8 ws = 1
A1 3 3 3 3 3 3
A2 1 1 1 1 1 1
A3 5 5 5 5 5 5
A4 6 6 6 6 6 6
A5 2 2 2 2 2 2
A6 4 4 4 4 4 4
The normalized decision matrix for VIKOR method can be obtained in a linear
method as shown in Table 5.6. Keeping in view with the normalized decision matrix,
* 
the best value ( f j ) and the worst value ( f j ) for the attributes are obtained using
*
equations 5.19(a) and 5.19(b) respectively. The best values and worst values are: f j

= (1, 1, 1, 1, 1, 1, 1, 1, 1, and 0.621051) and f j = (0.668445, 0.485991, 0.416653,
0.583383, 0.622911, 0.692054, 0.711443, 0.374821, 0.320459, 1). For ranking the
alternatives, the methodology needs to calculate Si, Ri along with the final values of
Qi as given in Table 5.15 based on equations 5.20, 5.21 and 5.22 where i=1, 2,….n. It
has been seen alternative A2 is best ranked by ‘minimum Q value’ and the stability in
decision making is completely satisfied (condition 2) for all weighing proportion but
the acceptance advantage (condition 1) is not satisfied as Q(A2)-Q(A5)=0.15<0.2.
Therefore, a final ranking of alternatives as shown in Table 5.15 is obtained through a
93
compromise solution satisfying equations 5.23 and 5.24 i.e. the alternative in the
second position (A5) forms a compromise solution together with the alternative (A2) in
the first position satisfying the conditions provided in VIKOR method. Considering
integrated weight with different proportion of objective and subjective weights, the
ranking of the alternatives is illustrated.
Table 5.15 The ranking and the compromise solutions
94
In PROMETHEE method, the decision maker gives his/her preference in order to
compare the alternatives for each attribute. A preference value ranging between 0
and 1 will be assigned to the ‘better’ alternative whereas the ‘worst’ alternative
receives a value 0. Based on this theory, a pair-wise comparison of attribute ‘depth of
seat’ is prepared as shown in Table 5.16. As ‘depth of seat’ is a beneficial attribute,
higher values are desired. Considering wo = 1 and ws = 0, the leaving (the measure of
the outranking character,   (A) (i.e. dominance of alternative A to other
alternatives)), entering (   (A) (i.e. degree to which alternative A is dominated by all

other alternatives) and net flows ( (A) ) are evaluated using equations 5.27, 5.28
and 5.29 respectively. The final ranking is illustrated in Table 5.17.
Table 5.16 Preference function(Pi) resulting from the pair wise comparisons of the six
alternatives with respect to criterion depth of cut.
A1 A2 A3 A4 A5 A6
A1 - 0 0 0 0 0
A2 1 - 1 1 1 1
A3 1 0 - 1 0 1
A4 1 0 0 - 0 1
A5 1 0 1 1 - 1
A6 0 0 0 0 0 -
   
Table 5.17 Positive   (A) , negative   (A) and net flows (A)  for the scenario
A1 A2 A3 A4 A5 A6   ( A)   ( A) (A) Rank
A1 - 0.213 0.213 0.391 0.213 0.210 1.24 2.967 -1.727 5
A2 0.787 - 0.696 0.843 0.517 0.968 3.811 0.995 2.816 1
A3 0.543 0.303 - 0.543 0.303 0.543 2.235 2.056 0.179 3
A4 0.608 0.125 0.276 - 0.193 0.432 1.634 2.912 -1.278 4
A5 0.787 0.354 0.514 0.775 - 0.968 3.398 1.226 2.172 2
A6 0.242 0 0.357 0.360 0 - 0.959 3.121 -2.162 6
Similarly, net flows for different proportion of objective and subjective weight
for all the attributes can be tried. The ranking thus obtained based on (A) value is
given in Table 5.18.
Table 5.18 Ranking of alternatives considering integrated weight
Alternatives wo = 1 wo = 0.8 wo = 0.6 wo = 0.4 wo = 0.2 wo = 0
ws = 0 ws = 0.2 ws = 0.4 ws = 0.6 ws = 0.8 ws = 1
A1 5 5 5 5 5 5
A2 1 1 1 1 1 1
A3 3 3 3 3 3 2
A4 4 4 4 4 4 4
A5 2 2 2 2 2 3
A6 6 6 6 6 6 6
95
The preference index values for different alternatives with respect to three MADM
methodologies are shown in Figures. 5.5, 5.6 and 5.7 for different weight proportion.
It is observed that alternate A2 is the best among all when attribute weight became
more subjective. The ranking order for alternatives changes according to the change
in proportion of attribute weight (subjective and objective). It has been found ranking
order for the alternatives change with increase of the proportion of objective weight
for all the methods. However, change of ranking order with increase of objective
weight is more pronounced in the VIKOR method. When only objective weight for
attributes is considered, A3 becomes best alternative instead of A2 in case of VIKOR.
In case of TOPSIS method, the ranking of alternative remains same whatever may
the weighing proportion. The final ranking of alternatives considering different
weighting proportion of objective and subjective weights is summarized in Table 5.19.
1.2
Preference index value
0.8
0.6 TOPSIS
VIKOR
0.4
PROMETHEE
0.2
0
A1 A2 A3 A4 A5 A6
Weighing proportion of subjective objective
value(wo=0 ws=1)
Figure 5.5 Integrated subjective objective weight with wo=0 and ws=1
96
1.2

1
0.8
0.6 TOPSIS
VIKOR
0.4
PROMETHEE
0.2
0
A1 A2 A3 A4 A5 A6
weight(wo=0.5 ws=0.5)
Figure 5.6 Integrated subjective objective weight with wo=0.5 and ws=0.5
1.2
0.8
0.6 TOPSIS
0.4 VIKOR
PROMETHEE
0.2
0
A1 A2 A3 A4 A5 A6
weight(wo=1 ws=0)
Figure 5.7 Integrated subjective objective weight with wo=1 and ws=0
97
Table 5.19 Ranking order comparison
98
5.4 Conclusions
In this research, an attempt has been made to select best office chair with ergonomic
considerations using three important MADM methodologies. In the selected MADM
approaches, attribute weights are determined using combination of objective and
subjective weights to emulate real life decision making process. It is observed that the
best alternative chosen remains same for different weighing proportions although the
selected MADM methods use different types of normalization to eliminate the units of
criterion functions and different ranking index measurement method. The proposed
method attempts to consider both subjective and objective weights of qualitative and
quantitative attributes and integrates them to decide the importance of weights of the
alternatives. Considering wo = 1 and ws = 0, the ranking of the alternatives is illustrated
for all the three methods such as TOPSIS, VIKOR, and PROMETHE. The alternatives
are arranged in the descending order of their preference as A2-A5-A1-A6-A3-A4. From the
above values of preference index, it is understood that the alternative designated as A2 is
the first right choice for the given design application under the given conditions
considering ten attributes of the product. It is observed that alternative 2 becomes the
best choice in all the three methods even if the weighting proportion of the objective and
subjective weight changes. It is also found that ranking order of alternatives 1, 3, 4 and 6
in PROMETHE changes with weighting proportion of objective and subjective weights. In
VIKOR method, preference order of alternatives 3 and 4 is altered when the proportion of
objective weight decreases. However, the second alternative is the best alternative in all
of the three methods whatever may be the proportion of weights of the attributes. The
result indicates that all MADM methods considered in this work behave in a similar
manner resulting same best alternative irrespective of proportion of weightings for
objective and subjective weights. Therefore, the decision makers have the liberty of
choosing the best method depending on ease of computational procedure. The method
uses only ten features of the product. In future, more features of the product can be
incorporated in the decision making process and other MADM approaches may be
explored.
99
CHAPTER 6
HUMAN UPPER ARM POSTURE

PREDICTION WITHIN
ISOCOMFORT WORK ZONE
6.1 Introduction
Posture analysis is an important issue for the performance analysis of tasks because
it is responsible for promoting health by minimizing stress and discomfort during work
(Haslegrave, 1994). Musculoskeletal disorders are frequently observed in work
environment during manual operation due to abnormal and poor working postures of the
body (Haslegrave, 1994; Westgaard and Aaras, 1984). However, it is difficult to identify
specific comfort posture (comfort link configuration) as the redundancy of human arm
develops many link configurations and joint motions to perform the same task (Gragg et
al., 2013). Extra degree of freedom (DOF) due to redundancy helps in free positioning
and moves around or between obstacles (Conkur and Buckingham, 1997; Chiaverini,
1997). As deviations from normal postures over a prolonged period of time results in
stress in joint muscles and other soft tissue muscles, it is important to design a
workspace to place all materials tools, and equipment within the work envelope so that
they are easily accessible by the operators (Grandjean and Hunting, 1977; Corlettet et
al., 1979, Corlett and Manenica, 1980; Das and Gardy, 1983). Improved layout of
workspace enables the operators to use their hands safely avoiding awkward postures
and thereby prevents the operators from serious injury problem (Kee, 2002). Therefore,
prediction of good posture within comfort workspace becomes a useful way for
enhancing the productivity by minimizing operator’s stress and injury (Lim and Hoffmann,
1997).
Most of the posture prediction model relies on analytical and heuristic approaches to
provide feasible postures with redundant degrees of freedom and infinite feasible
movements of arm (Jung et al., 1992; Kee et al., 1992). Although inverse kinematic (IK)
approach finds joint angle by considering the hand position in space but the complexity
of the IK solution increases with higher redundancy due to increase in number of DOF.
Increase in DOF leads to non-linear equations and singularity problem resulting in
indeterminate situation for posture (joint angles) evaluation. Since analytical, geometric,
iterative or algebraic method finds difficulty to provide complex IK solutions, the present
work proposes two artificial intelligence techniques to predict kinematic model of upper
arm posture with a comfort work envelop. Least Squares Support Sector Machines
(LSSVM) and Adaptive Neuro Fuzzy Inference System (ANFIS) enhance the model
capabilities by providing an infinite number of postures for highly redundant hand. A
kinematic model of human arm with seven degrees of freedom is considered to generate
a three dimensional workspace around operator. In order to avoid the physical
100
constraint, an isocomfort joint angle range (Diffrient et al., 1985) is considered to
determine the workspace. With the help of Indian anthropometric data for segment
lengths, a forward kinematics (FK) model is employed to develop a relation between
human comfort range of joint angles and the position of kinematic hand. Finally, a
workspace is generated with possible positions of kinematic hand. Once the positions
inside workspace (comfort work zone) obtained through FK solution, the model becomes
trained through LSSVM and ANFIS for estimating the IK solution of a 7-DOF kinematic
chain model (human arm model) to predict a comfort posture within the comfort work
zone.
6.2 Model description

Posture of human arm is predicted with a range of comfort joint angles and link
parameters. Human arm consists of three parts such as upper arm, lower arm and hand.
Although upper body can generate infinite postures with many degrees of freedom of
different parts, present work focuses only on arm for simplifying the model analysis. The
model considers a three link system of a kinematic chain with seven degrees of freedom
such as three glenohumeral joint, two elbow joints and two wrist joints.The glenohumeral
joint (also known as shoulder joint) is formed by the humeral head and the glenoid cavity
of the scapula. In general, it has three rotational degrees of freedom such as flexion and
extension, abduction and adduction and internal and external rotation. The elbow joint
can be regarded as a hinge joint with two degrees of freedom such as flexion-extension
and pronation-supination. Finally, wrist, a pivot joint, has two degrees of freedom like
flexion-extension and wrist deviation. Although the effect of spine develops discomfort
and fatigue, the spine and other parts are assumed to perform no joints movements. The
joints in the model are connected through links. Three links describe the segment
lengths such as upper arm, lower arm and hand. In order to carry out the kinematic
analysis, different segment measurements are selected from previous literature (Kaur et
al., 2011, Singh et al., 2013). The coordinate frames at each joint are defined by Denavit-
Hartenberg (D-H) convention. Four parameters (referred as D-H parameters) describe
the relative motion between two coordinate frames.
101
Kinematic analysis
Lc
Lua
Lfa
Lh
Figure 6.1 Human model arm representation(a) human arm(b)Coordinate frame
Figure 6.1 shows three coordinate frames correspnding to three joints with seven
degrees of freedom. For the present study, D-H notation is adopted to describe
kinematic model of human upper arm.
6.2.1 Denavit -Hartenberg Representation
D-H notation uses a 4 × 4 homogeneous transformation matrix representing each
link's coordinate system at the joint with respect to the adjacent link's coordinate system.
A kinematic chain model carries n joints (from 1 to n) with n  1 links (from 0 to n,
starting from base) and each joint is placed between two links. By this convention, joint i
connects link i  1 to link i. It is considered that the location of the joint i to be fixed with
respect to link i  1 . Each link of the kinematic chain model is rigidly attached to a
coordinate frame for performing the kinematics analysis.
102
Zi- Zi
Zi-
2
Joint
1 Joint
i-1 i+1
Link i- Lin
1 Joint i
ki
i
Y
i
Yi- Hi X
Oi- O
2 i
2
d Yi- a i
i
i
Xi- 1
2 1
Oi-
1 Xi-
1
Figure 6.2 The denavit Hartenberg notation (Mittal and Nagrath, 2007)
Figure 6.2 shows two length parameters ( a i and d i ) and two angle parameters (  i
and  i ). For a joint, the parameters, a i ,  i , and di are constant and determined by the
geometry of the link. a i is the distance between axes z i1 and zi (link length), d i is the
distance between axes x i1 and x i measured along axis z i1 (joint distance).  i is the
angle between axes z i1 and z i ,measured about axis x i (twist angle).  i is the angle
between axes x i1 and x i ,measured about axis z i1 (joint angle). As the joint moves,
only the parameter  i becomes the variable that represents the joint angular
displacement. With these four parameters, the transformation matrix, Tii1 can be
obtained showing the position and orientation of each coordinate frame with respect to
previous frame with its position and orientation.
An overview of all the parameters, used to describe the kinematic arm model, is
presented in Table 6.1.
103
Table 6.1 D-H parameters of human arm model
 d a 
Link 1     0 0 0
  2   1 
  
Link 2     0 0 

  2   2 
  
    2
Link 3     -lua 0 
     3   
 2   2
Link 4  4  0 0  
 
 2
Link 5  5  -Ifa 0 
 
2
Link 6     0 0 
    6   
 
2  2
Link 7  7  0 -Ih 
 
2
The segment lengths are collected from (Murray, 2004). Where, Lc, the length of
clavicle i.e. the distance from sternum to the shoulder, Lua, the length of upper arm, Lfa,
the length of lower arm and Lh, the length of hand. The joint angle range for different
comfort level (comfort zone) is derived from previous estimation (Diffrient et al., 1985)
and provided in Table 6.2.
Table 6.2 Joint angle range for comfort zone (Diffrient et al., 1985)
Joint Posture Comfort Zone different comfort range
Shoulder extension-flexion -150—350
Shoulder adduction-abduction -250-00
Shoulder rotation -200-450
Elbow flexion 150-1000
Elbow supination-pronation -900—300
Wrist ulnar radial deviation -150-50
Wrist extension-flexion -250-450
D-H notation of the joint is introduced with some convention to solve this matrix. The
convention and steps for D-H notation is presented as follows.The following steps based
on D-H notation are used for deriving the forward kinematics.
Step 1: Base frame is assigned. Set the origin anywhere on the z 0  axis . The x 0 and y 0
axes are chosen conveniently to form a right-hand frame.

Step 2: The origin o i is located, where the common normal to z i and z i1 intersects at z i .
If z i intersects z i1 , a i located at this intersection. If z i and z i1 are parallel, locate o i in
any convenient position along z i .
104
Step 3: x i is considered along the common normal between z i1 and z i through o i , or in
the direction normal to z i1  z i plane if z i1 and z i intersect.
Step 4: y i is established to complete a right-hand frame.
Step 5: The end-effector frame is assigned as on x n y n z n . Assuming the nth joint is
revolute, set z n  a along the direction z n1 . The origin o n is taken conveniently along z n
direction, preferably at the centre of the gripper or at the tip of any tool that the
manipulator may be carrying.
Step 6: All the link parameters i , ai , di ,  i are tabulated.
Step 7: The homogeneous transformation matrices A i is determined by substituting the

parameters from table 6.1 in equation 6.1.
0
Step 8: Then the global transformation matrix TEnd is formed using equation 6.2. This
then gives the position and orientation of the frame expressed in base coordinates.
In this convention, each homogeneous transformation matrixis A i represented as a
product of four basic transformations:
cos i   sin i cos i  sin i  sin i  a i cos i 

 sin  cos i cos i   cos i  sin i  a i sin i   (6.1)
Ai   i
 0 sin i  cos i  di 
 
 0 0 0 l 
Where four quantities i , ai , di , i are parameter associated with link i and joint j. The
four parameters i , ai , di , i in the above equation are generally given name as joint
angle, link length, link offset, and link twist respectively.By substituting the D-H
parameters from Table 6.2 in equation 6.1, the individual transformation matrices A1 to
0
AEnd can be obtained and the global transformation matrix ( TEnd ) from the first joint to
the last joint of the 7-DOF Redundant manipulator can be derived by multiplying all the
individual transformation matrices. So,
nx ox ax px 
n oy ay p y  (6.2)
0
TEnd  A1A 2 A 3 A 4 A 5 A 6 A 7 A End   y
nz oz az pz 
 
0 0 0 1
105

Where, p x ,p y ,p z are the positions and (nx ,ny ,nz ),(ox , oy , oz ), and (ax , ay , az ) are the 
orientations of the end-effector. The orientation and position of the end-effector can be
calculated in terms of joint angles and the D-H parameters of the manipulator are shown
in following equations:
n x  c 7 c 6 c 5  s 7 s 5 c 3 c 4 c 1c 2  s 1s 2   s 4 c 1s 2  s 1c 2   c 7 s 5 c 6  s 7 c 5 s 3 s 1s 2  c 1c 2 
 s 6 c 7 c 3 s 4 s 1s 2  c 1c 2   c 4 c 1s 2  s 1c 2 
(6.3)
n y  c 7c 6c 5  s7 s5 c 3c 4 s1c 2  c 1s 2   s 4 c 1c 2  s1s 2   c 7 s5c 6  s7c 5  s3 s1c 2  c 1s 2 

 s 6c 7  c 3 s 4 s1c 2  c 1s 2   c 4 c 1c 2  s1s 2 
(6.4)
nz  c 7c 6 s3 c 4 s5  c 7c 6 s5c 3  c 7 s6 s3c 3  c 7 s6 s3 s 4  s7 c 3c 5  s7 s3c 4 s5

(6.5)
o x  c 6 c 3 s 4 s1s 2  c 1c 2   c 4 c 1s 2  s1c 2   s 6 c 5 c 3 c 4 c 1c 2  s1s 2   s 4 c 1s 2  s1c 2 

 s 6 s 5 s 3 s1s 2  c 1c 2  (6.6)
o y  c 6  c 3 s 4 s1c 2  c 1s 2   c 4 c 1c 2  s1s 2   s 6 c 5 c 3 c 4 s1c 2  c 1s 2   s 4 c 1c 2  s1s 2  (6.7)

 s 6 s 5  s 3 s1c 2  c 1s 2 
o z  s6 s3c 4c 5  s5c 3 s6  s3c 4c 6

(6.8)
a x  s 7 c 6 c 5  c 7 s 5 c 3 c 4 c 1c 2  s1s 2   s 4 c 1s 2  s1c 2   s 7 s 5 c 6  c 7 c 5 s 3 s1s 2  c 1c 2 

 s 6 s 7 c 3 s 4 s1s 2  c 1c 2   c 4 c 1s 2  s1c 2 
(6.9)
a y  s7c 6c 5  c 7s5 c 3c 4 s1c 2  c1s2   s4 c1c 2  s1s2   s7c 6s5  c 7c 5  s3 s1c 2  c1s2 
 s7s6  c 3s4 s1c 2  c1s2   c 4 c1c 2  s1s2 
(6.10)
a z  s7 c 6 s3 c 4c 5  s7c 6 s5c 3  s7 s6 s 4 s3  c 7 c 4 s3 s5  c 3 c 5 c 7
(6.11)
p x  d 7 s 7 c 6 c 5  c 7 s 5 c 3 c 4 c 1c 2  s1s 2   s 4 c 1s 2  s1c 2   d 7 s 7 s 5 c 6  c 7 c 5 s 3 s1s 2  c 1c 2 

 d 7 s 7 s 6  d 5 c 3 s 4 s1s 2  c 1c 2   c 4 c 1s 2  s1c 2    d 3 c 1s 2  s1c 2 
(6.12)
106
p y  d7 s 7 c 6 c 5  s 5 c 7 c 3 c 4 s1c 2  c 1s 2   s 4 c 1c 2  s1s 2   d7 s 7 s 5 c 6  c 7 c 5  s 3 s1c 2  c 1s 2 
 d7 s 7 s 6  d5  c 3 s 4 s1c 2  c 1s 2   c 4 c 1c 2  s1s 2 
(6.13)
p z  d7 s7 s 3 c 6 c 5 c 4  d7 s7 s 5 c 6 c 3  d7 s 7 s 6 s 4 s 3  d7 s 5 s 3 c 7 c 4  d7 c 7 c 5 c 3  d5 s 4 s 3  d1
(6.14)
where c i  cos(i ), si  sin(i ) . From equation (6.3)-(6.14), the position and orientation
of the 7-DOF redundant manipulator can be obtained and the exact value of these
equations can be calculated if all the joint angles and link parameters are given. This is
the solution to the forward kinematics.
6.3 Methods for determination of IK solutions
This chapter investigates the use of two artificial intelligence techniques to produce
the solution to the inverse kinematics problem for a three joints upper arm kinematic
chain.
6.3.1 LSSVM Architecture
The basic formulation of the standard LSSVM (Suykens and Vandewalle, 1999). for
function estimation is briefly described in this section. Consider a given training set of N
data points x k , y k Nk 1 with input data x k  RN and output y k  r , where R N the N-
dimensional vector space and r is is the one-dimensional vector space. This study uses
position in workspace as input (x) parameters of the LSSVM. The output of LSSVM is
joint angle(y).
Following regression model is used.
y( x)  w T ( x)  b , (6.15)
Where the nonlinear mapping (.) maps the input data into a higher dimensional feature
space; w  Rn ; b  r ; w=an adjustable weight vector; b=the scalar threshold,

The following optimization problem is formulated:
N
Minimize: 1 w T w  1   e k2 (6.16)
2 2 t 1
subject to: y( x)  W T x k   b  ek k=1,……..,N (6.17)

where, et is the error variable at time t, where ( ) is a nonlinear function mapping the
input space into a higher dimensional space and  is the regulation constant. The
Lagrange function can be obtained as
107
1 N
 
N
LW, b, e,    w w    e i2    i w T x i   b  e i  y i
1 T (6.18)
2 2 i1 t 1
where  t are the Lagrange multipliers. The solution of the above Eq.(6.18) can be
obtained by partially differentiating with respect to each variable

L N
 0  w    K ( x k )
w k 1
L N
 0   K  0
b k 1
L k=1,………,N
 0   k  e k
e k
L
 0  w T x k   b  e k  y k  0 k=1,……..,N (6.19)
 K
When the variables w and e are removed, the equation can be rewritten as a linear
function group
0 1T   b  0 (6.20)
     
1N    1I    y 
 
Where y=y1 …yN,   1,......... .. N and Mercer’s theorem (Smola et al. 1998; Vapnik,
1998), is applied within the  matrix,
  x k  x l   kx k , x l  ,

T
k, l=1,…….N
Where k (xk, xl) is the kernel function. Choosing   0 , ensures the matrix
0 1T   and b is given by
 1 
is invertible. Then the analytical of
1    I
 x  x l x k  x l 
 
T

K x k , x l   exp  k  k,l=1,……………N (6.21)

 2 2 

Where  is the width of radial basis function.
The resulting LSSVM model for joint angle prediction becomes then
N
Joint angle=   K k x, x k   b (6.22)
k 1
The above described LSSVM has been adopted for prediction of joint angle
6.3.2 ANFIS Architecture
Previous chapter (chapter 3) presents the relevant methodology of ANFIS
architecture. During training, a five layered ANFIS structure is constructed with one input,
three hidden and one output. The Gaussian type of membership function (gaussmf) is
108
used for input and linear type function is used for output. The number of correct outputs
is noted till the error is minimized.

With the help of comfort joint angle range as comfort posture and segment length,
the kinematic equations are solved and hence trajectory can be achieved by moving
each joint gradually to the determined position. A comfort work zone can be created by
using equation 6.12, 6.13 and 6.14. Figure 6.4 shows the comfort work zone with hand
reach positions (px,py and pz).
Figure 6.3 Workspace for 7-DOF redundant manipulator showing human arm extremity
As it is difficult to solve the nonlinear equation with the values of hand reach
position(px,py and pz) in order to find out the comfort joint angles( 1 ,  2 ,  3 ,  4 ,  5 ,  6
and  7 ), in the proposed approach consider LSSVM and ANFIS model to obtain an IK
solutions. Hand reach position that obtained from the forward kinematic relations are
considered as input for LSSVM and ANFIS model and posture of upper arm in terms of
joint angle (extracted from literature) as output in order to predict an improved set of joint
angles. A sample of input-output data are trained in order to predict 1 ,  2 ,  3 ,  4 ,  5 ,  6
and  7 . Each of the networks carries seven different models with input px (input1), py
109
(input2) and pz. (input3) and output 1 ,  2 ,  3 ,  4 ,  5 ,  6 and  7 respectively. In order to
construct the model, dataset of 1600 training data and 400 data testing data was
considered. Both ANFIS and LSSVM networks will be trained with position px,py and pz
as inputs and corresponding joint angles theta1, theta2, theta3, theta4, theta5, theta6
and theta7values as output. The matrix data1 contains the px,py and pz- 1 values to train
the first network. Similarly, for second network will be trained with all position values as
input and corresponding  2 value as output and so on. It has been seen that the
predicted joint angle falls within the isocomfort joint angle range. It has been seen from
table 6.3 that in both the model the root mean square error have potential values that
indicate that this solution is accurate for individual data and may be useful for future
posture prediction system. Both model are able to map Cartesian coordinates of position
point on comfort zone to healthy biomechanical configurations(joint angles).The surface
plot in case of ANFIS as shown in figure.6.5 shows an uniform distribution of data and
homogeneity in the training data. Therefore it has been confirmed that the solution is
adequate for predicting a comfort posture with a healthy trajectory around the operator.
Low residual value of LSSVM in comparison to ANFIS shows that LSSVM provides a
better solution in comparison to ANFIS. In this study, the radial basis function is used as
the kernel function of LSSVM. Figure 6.6 shows the residual for all seven angles which
are nothing but the difference between the predicted output from the model and the
actual output of joint angles obtained from literature. As the points are randomly spread
around the horizontal axis, the prediction is found to be appropriate. Figure shows that
the residual obtained from LSSVM model (red colour) are very close to the horizontal
axis in comparison to the blue lines for ANFIS model which indicates LSSVM model
shows less error in comparison to ANFIS model.
110
1 2
3 4
5 6
111
7
Figure 6.4 Surface plot for 1 ,  2 ,  3 ,  4 ,  5 ,  6 ,  7
Table 6.3 Prediction comparison between LSSVM and ANFIS

ANFIS LSSVM
Training (RMSE) Testing Training Testing

(RMSE) (RMSE) (RMSE)
1 0.006416 0.335284 0.000156 0.067134
2 0.023065 0.20086 0.001387 0.015626
3 0.006504 0.555197 0.003414 0.033452
4 0.00158 0.591968 0.000818 0.234112
5 0.002328 0.418890 0.000899 0.095139
6 0.010064 1.14556 0.008165 0.332050
2 0.001673 0.2986661 0.000891 0.043887
112
θ1 LSSVM θ2 LSSVM
0.0003 ANFIS 0.0002 ANFIS
0.00015
0.0002
0.0001
0.0001
Residual
0.00005
Residual
0 0
0 1000 2000 -0.00005 0 1000 2000
-0.0001
-0.0001
-0.0002
-0.00015
-0.0003 -0.0002
Number of observations Number of observations
θ3 LSSVM θ4 LSSVM
0.0002 ANFIS
ANFIS
0.00015 0.0015
0.0001 0.001
Residual
0.00005
0.0005
Residual
0
0
-0.00005 0 1000 2000
-0.0001 -0.0005 0 1000 2000
-0.00015 -0.001
-0.0002 -0.0015
θ5 LSSVM θ6 LSSVM
ANFIS ANFIS
0.0003
0.0003
0.0002 0.0002
0.0001 0.0001
Residual
Residual
0
0
0 500 1000 1500 2000 -0.0001 0 1000 2000
-0.0001
-0.0002
-0.0002 -0.0003
-0.0003 -0.0004
113
θ7 LSSVM
ANFIS
0.0003
0.0002
0.0001
Residual
0
0 1000 2000
-0.0001
-0.0002
-0.0003
-0.0004
Number of observations
Figure 6.5 Residual plot for 1 ,  2 ,  3 ,  4 ,  5 ,  6 ,  7 through LSSVM and ANFIS
To ensure effective performance, the operator needs to work within a comfort work
zone so that there is no interference during manual operation. Comfort posture also
provides a healthy and safety work environment. In this study, an isocomfort joint posture
has been considered from previous literature to create a comfort work zone. With the
help of Indian anthropometric data and D-H notation, the proposed model helps to
analytically generate three-dimensional isocomfort work zone by using forward
kinematics for a range of isocomfort postures. Two intelligence techniques have been
applied to predict comfort joint arm postures considering end position of hand in
workspace.
The results of this study may be generalized or justified to a larger extent by
considering more degrees of freedom resulting in complex situation to the link system.
Therefore, only the upper extremity is considered for the analysis. The results show that
the predicted joint angles are found within the isocomfort range. No significant difference
between actual and predicted joint is observed in both the techniques as the root mean
square error has low values. Table 6.4 shows LSSVM and ANFIS based root mean
square error (RMSE) results.
6.5 Conclusions
This study predicts the comfort joint angles for human arm in sitting and standing
postures. The study develops an inverse kinematics solution for 7-DOF kinematic chain
model using ANFIS and LSSVM. The difference in actual and predicted joint angle with
ANFIS and LSSVM model for 7-DOF kinematic chain model clearly depicts that the
proposed LSSVM method performs better as minimum RMSE error is observed. The
114
model is constructed considering hand end position as input and seven joint positions as
output parameters in training and testing data with a smaller number of iteration steps.
Hence, it is concluded that the trained LSSVM and ANFIS models can be utilized to
solve complex, nonlinear and discontinuous kinematics equation of complex kinematic
chain model for biomechanical studies in order to predict comfort work zone.
115
CHAPTER 7
EXECUTIVE SUMMARY
AND CONCLUSIONS
7.1 Introduction
The present thesis involves detailed study on affective satisfaction of user, product’s
functional performance and comfort of the user to enhance the design process of
product/machine/equipment/component. Affective satisfaction of user deals with the
issues related to product’s intangible characteristics that satisfy the users. Objective
performance originates from the effectiveness of use whereas the comfort level of user is
quantified through biomechanical and physiological perspective of human body during
physical interaction with the product. Consideration of three criteria such as product
performance in terms of usability, user satisfaction and comfort level generally enhances
the design process. To deal with the complex operator-system interaction, the study
outlines various approaches relevant to design phase of a product. The study also
emphasizes on prediction of comfort level through biomechanical analysis of human
body-product interaction. In order to meet the objectives, the methodology proposed
here as follows:
1. An integrated approach is established that deals with subjective and objective
design criteria of product with ergonomic consideration. In order to deal with
subjective feelings associated with product, a questionnaire survey has been
conducted. Customer driven approaches like QFD and factor analysis are
considered in order to establish the relation between user requirements and
design elements. Intelligence technique like ANFIS develops a relation between
design parameters and customer satisfaction. The approach is described with the
help of an office chair design.
2. The study also involves a biomechanical analysis as a measure of comfortness
during physical interaction of user and product. To describe the biomechanical
analysis, human-chair seat model with various parameters is considered as a
case study. Present work develops a simple two-dimensional finite element
model of human soft tissue with ischial tuberocity-seat with various seat material
parameters, thicknesses of seat and different loading angle to predict stress at
ischial tuberosity in order to provide guidelines to reduce occurrence of pressure
ulcer. The analysis investigates maximum extent of stress in soft tissue muscle
(at ischial tuberocity) on prolonged sitting in an office environment.
3. The study carries out various possible ways to choose a suitable ergonomically
design product with most usability factors (design criteria) among a number of
conflicting criteria.
116
4. As awkward posture is always associated with a measure of discomfort, the
analysis covers the prediction of comfort posture with design of comfort work
zone in order to place all materials, tools and equipment within the work envelope
so that they are easily accessible by the operators.
7.2 Summary of findings
 A novel integrated approach using statistical and artificial intelligence techniques
has been proposed in this thesis to handle effectively subjective and objective
design characteristics. Customers’ expectation from the product is extracted
through a questionnaire survey. Data reduction technique like factor analysis has
been applied to survey data to eliminate redundancy. The reduced customer
requirements are translated into design characteristics using QFD. Through an
artificial intelligence technique like adaptive neuro-fuzzy inference system, the
nonlinear relationship between user satisfaction and design attributes can be
successfully established. The prototype is compared with the data prescribed by
Bureau of Indian Standard (BIS) for office chair. It is observed that some
parameters are not within the range of BIS data. The variations are attributed to
localization of sample data.
 A numerical approach based on simple two dimensional finite element seat-soft
tissue model is proposed in the thesis to provide guidelines for the designers for
analyzing the behavior of interaction of soft human tissue and cushion material
and estimates maximum stress beneath the bony structure (ischial tuberocity). It
can be deduced that the size of the affected zone as well as the stress is much
larger for a rigid seat as compared to soft cushion. The cushion having elastic
modulus of 20 kPa and density 40 kg/m3 shows less stress distribution at ischial
tuberosity in comparison to cushion having elastic modulus of 200 kPa and
density 60 kg/m3. The effect of sitting posture (tilting posture) from 0 0 to 30 0 has
also been analyzed to study the influence of postural change on maximum stress
at ischial tuberocity. It has been observed that only a fraction of total vertical load
acts on ischial tuberosity due to postural change because change of posture
transfers the load to those regions of body other than ischial tuberosity.
Therefore, the muscles lying below ischial tuberosity relieves stress by
distributing the stress toward thighs and waist. The stress at ischial tuberosity
goes on increasing with sitting time. However, the stress becomes constant after
117
an interval of one and half hours of continuous sitting but the intensity of load
causes significant influence on cell damage.
 The proposed MADM method in this work considers both subjective and objective
weights of design attributes for selection of alternatives in an uncertain
environment so that the decision maker facilitate with the objective information
regarding the product as well as the uncertainity of human judgement on the
product. In order to rank the alternatives with different weighting proportion, three
MADM methods like TOPSIS, VIKOR and PROMETHEE have been used. In
order to check the stability of ranking with respect to different weighted criteria, a
sensitivity analysis has been carried out considering different proportion of
attribute weights (subjective and objective). It is to be noted that the alternative
designated as 2 is the first preference by all the methods even if the weighting
proportion of subjective and objective weights vary.
 The proposed kinematic model of human arm makes it possible to evaluate
analytically comfort work zone satisfying a comfort posture. Human arm with
seven degrees of freedom is considered to generate a three dimensional
workspace around operator. In order to avoid the physical constraint, an
isocomfort joint angle range (Diffrientet al., 1985) is considered to determine the
workspace. As it is difficult to solve complex inverse kinematic equations, this
chapter explores two artificial intelligence techniques such as LSSVM and ANFIS
model to predict upper arm comfort posture under a standing/sitting condition
satisfying comfort work zone.
7.3 Contribution of research work
 Providing a novel integrated approach using statistical and artificial intelligence
techniques, current methodology manages subjective and objective design
criteria for product development with ergonomic consideration and provides a
guideline to adopt this approach in any design phase of any product. Through an
artificial intelligence technique like adaptive neuro fuzzy inference system, the
model is able to develop a nonlinear relationship between customer satisfaction
and design attributes.
 As the proposed design does not satisfy all prescribed limits regarding design
elements of Bureau of Indian Standard (BIS), it has been suggested that the
variations are attributed to localization of data collection. The standards should
118
be regularly reviewed and formulated in accordance with the technological
development as well as localization effect of anthropometric parameters.
 The numerical analysis considering simple 2D finite element formulation provides
insight into the problem and prescribes guidelines to avoid suffering from
pressure sore to some extent. It also suggests the ways to reduce the stress on
bony prominence causing cell death of muscle tissue during prolonged sitting in
an office environment.
 Effect of changing posture suggests that change in posture transfers the load to
those regions of body other than ischial tuberocity and relieves the muscles from
the load below ischial tuberocity by distributing the stress towards thighs and
waist. In order to make stress relaxed or reduce the direct load on fibers, either a
postural change or a leisure time should be preferred after certain duration of
sitting. It has been shown that use of right kind of foam for seat cushion and
thickness can substantially reduce the stress level at ischial tuberosity.
 A novel multi attribute decision making (MADM) approach is proposed in the
present work to choose a suitable alternative (office chair) considering both
subjective and objective weights for design attributes in an uncertain
environment.
 Sensitivity analysis shows that different weighting proportion for subjective and
objective weights significantly influence in choosing the alternatives. It is to be
noted that the alternative designated as 2 is the first preference by all the
methods even if the weighting proportion of subjective and objective weights
varies. The model demonstrates that the ranking of alternatives depends upon
the weighting proportion of subjective and objective weights. However, different
approaches lead to approximately same alternatives preference order for a given
weighting proportions.
 The kinematic analysis of upper extremities allows the operator to have comfort
work zone within which possible posture can be accepted to enable the operator
to use their hands safely to perform the task. The proposed LSSVM and ANFIS
models can produce solution to complex inverse kinematics problem to suggest
the joint angles to reach at specified locations within a safe work zone.
119
7.4 Limitations of study
 The work has been limited to analyze only one performance i.e. sitting in office
environment.
 Although numerical analysis shows the stress within muscle decreases with
proper cushion thickness and postural changes, this work considers a simple 2D
model to provide guidelines for the designers and analyze behavior of interaction
of soft human tissue and cushion material. The study can be improved
considering a model with 3D formulation.
 Contouring of the seat can potentially influence on pressure distribution at ischial
tuberosity but this effect has not been considered in the present work.
7.5 Scope of future research
 The work can be extended to design of hand tools, machinery, vehicles and
furniture with ergonomic consideration used in various work environments.
 The numerical approach considered in this work may be extended to dynamic
analysis where vibrational effect can be analyzed in a moving vehicle.
 .Contouring of the seat can be considered to study its influence on pressure
distribution at ischial tuberosity.
120
Bibliography
1. Abdel-Malek, K., Yu, W., Yang, J. and Nebel, K. (2004). A mathematical
method for ergonomic-based design: placement. International Journal of
Industrial Ergonomics, 34(5), 375-394.
2. Anderson, G. B. J., Murphy, R. W., Örtengren, R. and Nachemson, A. L.
(1979). The influence of backrest inclination and lumbar support on lumbar
lordosis. Spine, 4(1), 52-58.
3. Andreas, K., Mutlu, B., Forlizzi, J., Guastrin C. and Hodgins, J. (2007). Robust,
low cost, non-intrusive sensing and recognition of seated postures.
Proceedings of the 20th Annual ACM Symposium on User Interface Software
and Technology, New Port, Rhode Island, USA, 7–10 October,149–158.
4. Armstrong, L. E., De Luca, J.P. and Hubbard, R. W. (1990).Time course of
recovery and heat acclimation ability of prior exertional heat stroke patients.
Medicine and Science in Sports and Exercise, 22(1), 36-48.
5. Avalle, M., Belingardi, G. and Montanini, R. (2001). Characterization of
polymeric structural foams under compressive impact loading by means of
energy-absorption diagram. International Journal of Impact Engineering, 25(5),
455-472.
6. Axelsson, J. R. C. (2000). Quality and ergonomics-towards successful
integration. Doctoral Thesis, Dissertation No 616, Quality and Human Systems
Engineering, Linkoping University, Linkoping.
7. Bellman, R. E. and Zadeh, L. A. (1970). Decision-making in a fuzzy
environment. Management Science, 17(4), 141–164.
8. Bennett, J. L. (1984). Managing to meet usability requirements: establishing
and meeting software development goals. Visual display terminals, Prentice-
Hall, Englewood Cliffs, NJ, 161-184.
9. Bergqvist, U. O. and Knave, B. G. (1994). Eye discomfort and work with visual
display terminals. Scandinavian Journal of Work, Environment and Health,20,
27-33.
10. Besharati, B., Luo, L., Azarm, S. and Kannan, P. K. (2006). Multi-objective
single product robust optimization: An integrated design and marketing
approach. Journal of Mechanical Design, 128(4), 884-892.
121
11. Bolstad, G., Benum, B. and Rokne, A. (2001). Anthropometry of Norwegian
light industry and office workers. Applied Ergonomics, 32(3), 239-246.
12. Brans J. P., Mareschal, B. and Vincke, P. (1984). PROMETHEE: a new family of
outranking methods in multi criteria analysis. Proceedings of Operational
Research, Amsterdam, North Holland, 477-490.
13. Braun, W. J., Rebollar, R. and Schiller, E. F. (1996). Computer aided planning
and design of manual assembly systems. International Journal of Production
Research, 34(8), 2317-2333.
14. Bridger, R.S. (1995). Introduction to Ergonomics. McGraw-Hill, New York.
15. Briody, C., Duignan, B., Jerrams, S. and Tiernan, J. (2012). The implementation
of a visco-hyperelastic numerical material model for simulating the behaviour of
polymer foam materials. Computational Materials Science, 64, 47-51.
16. Buckle, P. W. and Devereux J. J. (2002). The nature of work-related neck and
upper limb musculoskeletal disorders. Applied Ergonomics, 33, 207-217
17. Burri Jr, G. J. and Helander, M. G. (1991). A field study of productivity
improvements in the manufacturing of circuit boards. International Journal of
18. Carlsson, C. and Fullér, R. (1996). Fuzzy multiple criteria decision making:
Recent developments. Fuzzy Sets and Systems, 78(2), 139-153.
19. Castellucci, H. I., Arezes, P. M. and Viviani, C. A. (2010). Mismatch between
classroom furniture and anthropometric measures in Chilean schools. Applied
Ergonomics, 41(4), 563-568.
20. Ceelen, K. K., Stekelenburg, A., Mulders, J. L. J., Strijkers, G. J., Baaijens, F.
P. T., Nicolay, K. and Oomens, C. W. J. (2008). Validation of a numerical model
of skeletal muscle compression with MR tagging: a contribution to pressure
ulcer research. Journal of Biomechanical Engineering, 130(6), 061015(1-8).
21. Chaffin. D.B. 1999 Manual material-handling limits. In: Chaffin. DB. Andersson.
G.B.J. Martin. B.J. (Eds.). Occupational Biomechanics. 3rd Edition. Wiley. New
York, 315-354.
22. Chan, K. Y., Dillon, T. S., Kwong, C. K. and Ling, S. H. (2011, June). Using
genetic programming for developing relationship between engineering
characteristics and customer requirements in new products. IEEE International
Conference of Industrial Electronics and Application (ICIEA 2011), Beijing,
China, 526–531.
122
23. Chen, C. T. (2000). Extensions of the TOPSIS for group decision-making under
fuzzy environment. Fuzzy sets and systems, 114(1), 1-9.
24. Chen, C. T., Lin, C. T. and Huang, S. F. (2006). A fuzzy approach for supplier
evaluation and selection in supply chain management. International Journal of
Production Economics, 102(2), 289-301.
25. Chen, L. H. and Ko, W. C. (2008). A fuzzy nonlinear model for quality function
deployment considering Kano’s concept. Mathematical and Computer
Modeling, 48(3), 581-593.
26. Chen, S. H. (1985). Ranking fuzzy numbers with maximizing set and minimizing
set. Fuzzy sets and Systems, 17(2), 113-129.
27. Chen, S. J. J., Hwang, C. L., Beckmann, M. J. and Krelle, W. (1992). Fuzzy
multiple attribute decision making: methods and applications. Springer-Verlag
New York, Inc.
28. Chiaverini, S. (1997). Singularity-robust task-priority redundancy resolution for
real-time kinematic control of robot manipulators. Robotics and Automation,
IEEE Transactions on Systems, 13(3), 398-410.
29. Chow, W. W. and Odell, E. I. (1978). Deformations and stresses in soft body
tissues of a sitting person. Journal of Biomechanical Engineering, 100(2),79-87.
30. Chuang, M. C., Chang, C. C. and Hsu, S. H. (2001). Perceptual factors
underlying user preferences toward product form of mobile
phones. International Journal of Industrial Ergonomics, 27(4), 247-258.
31. Cimino, A., Longo, F. and Mirabelli, G. (2009). A multi measure-based
methodology for the ergonomic effective design of manufacturing system
workstations. International Journal of Industrial Ergonomics, 39(2), 447–455.
32. Conkur, E. S. and Buckingham, R. (1997). Clarifying the definition of
redundancy as used in robotics. Robotica, 15(5), 583-586.
33. Corlett, E. N., Madeley, S. J. and Manenica, I. (1979). Posture targetting: a
technique for recording work postures. Ergonomics, 22 (3), 357-366.
34. Corlett, E. N. and Manenica, I. (1980). The effects and measurement of working
postures. Applied Ergonomics, 11(1), 7-16.
35. Coulibaly, S., and Hua, Z. (2006). A scientific basis for the choice of scale
in QFD. International Journal of Services and Operations Management, 2(2),
124-145.
123
36. Crenshaw, R. P. and Vistnes, L. M. (1987). A decade of pressure sore
research. Journal of Rehabilitation Research and Development, 26(1), 63-74.
37. Dandavate, U., Sanders, E. B. N. and Stuart, S. (1996, October). Emotions
matter: user empathy in the product development process. In Proceedings of
the Human Factors and Ergonomics Society Annual Meeting, 40(7), 415-418).
SAGE Publications.
38. Das, B. and Grady, R. M. (1983). Industrial workplace layout designs: an
application of engineering anthropometry. Ergonomics, 26(5), 433-447.
39. Das, B. and Sengupta, A. K. (1996). Industrial workstation design: a systematic
ergonomics approach. Applied Ergonomics, 27(2), 157-163.
40. Dewangan, K. N., Owary, C. and Datta, R. K. (2010). Anthropometry of male
agricultural workers of north-eastern India and its use in design of agricultural
tools and equipment. International Journal of Industrial Ergonomics, 40(5), 560-
573.
41. Diffrient, N., Tilley, A. R. and Harman, D. (1985).Human scale (Cambridge, MA:
MIT Press).
42. Santos, D., I. J. A. L., Farias, M. S., Monteiro, B. G., Falcão, M. A. and
Marcelino, F. D. (2011). Using participatory ergonomics to improve nuclear
equipment design. Journal of Loss Prevention in the Process Industries, 24(5),
594-600.
43. Dursun Kaya, M., Samet Hasiloglu, A., Bayramoglu, M., Yesilyurt, H. and Fahri
Ozok, A. (2003). A new approach to estimate anthropometric measurements by
adaptive neuro-fuzzy inference system. International Journal of Industrial
Ergonomics, 32(2), 105-114.
44. Ellegast, R. P., Kraft, K., Groenesteijn, L., Krause, F., Berger, H. and Vink, P.
(2012). Comparison of four specific dynamic office chairs with a conventional
office chair: impact upon muscle activation, physical activity and posture.
Applied Ergonomics, 43(2), 296-307.
45. Fathallah, F. A. (2010). Musculoskeletal disorders in labor-intensive agriculture.
46. Falck, A. C., Örtengren, R. and Högberg, D. (2010). The impact of poor
assembly ergonomics on product quality: a cost–benefit analysis in car
manufacturing. Human Factors and Ergonomics in Manufacturing & Service
Industries, 20(1), 24-41.
124
47. Ferguson-Pell, M.W. (1990). Seat cushions selection. Journal of Rehabilitation
Research and Development, 2, 49-73.
48. Fredriksson, K., Bildt, C., Hägg, G. and Kilbom, Å. (2001). The impact on
musculoskeletal disorders of changing physical and psychosocial work
environment conditions in the automobile industry. International Journal of
49. Fok, K. S. and Chou, S. M. (2010). Development of a finger biomechanical
model and its considerations. Journal of Biomechanics, 43(4), 701-713.
50. Gauvali, M.K. and Boudolos, K. (2006). Match between school furniture
dimensions and children’s anthropometry. Applied Ergonomics, 37(6), 765–773.
51. Gefen, A., Gefen, N., Linder-Ganz, E. and Margulies, S. S. (2005). In vivo
muscle stiffening under bone compression promotes deep pressure sores.
Journal of Biomechanical Engineering, 127(3), 512-524.
52. Gefen, A., van Nierop, B., Bader, D. L. and Oomens, C. W. (2008). Strain-time
cell-death threshold for skeletal muscle in a tissue-engineered model system for
deep tissue injury. Journal of Biomechanics, 41(9), 2003-2012.
53. Gibson, R.F. (2012). Principle of Composite Material Mechanics, third edition.
CRC Press, University of Nevada, Reno, USA.
54. Goonetilleke, R.S. and Feizhou, S. (2001). A methodology to determine
the optimum seat depth. International Journal of Industrial Ergonomics. 27(4),
207–217.
55. Gragg, J., Yang, J. J. and Howard, B. (2012). Hybrid method for driver
accommodation using optimization-based digital human models. Computer-
Aided Design, 44(1), 29-39.
56. Grandjean, E. and Hünting, W. (1977). Ergonomics of posture-review of various
problems of standing and sitting posture. Applied Ergonomics, 8(3), 135-140.
57. Groenesteijn, L., Vink, P., Loozc, M.D. and Krause, F. (2009). Effects of
differences in office chair controls, seat and back rest angle, design in relation
to tasks. Applied Ergonomics, 40(3), 362–370.
58. Grujicic, M., Pandurangan, B., Arakere, G., Bell, W. C., He, T. and Xie, X.
(2009). Seat-cushion and soft-tissue material modeling and a finite element
investigation of the seating comfort for passenger-vehicle occupants. Materials
and Design, 30(10), 4273-4285.
125
59. Guan, S. S. and Lin, Y. C. (2001). A study on the color and style collocation of
mobile phones using neural network method. Journal of the Chinese Institute of
Industrial Engineers, 18(6), 84-94.
60. Hair, J.F., Black, W.C., Babin, B.J. and Anderson, R.E. (2010) Multivariate Data
Analysis, 7th ed., John Wiley and Sons, New York.
61. Hales, T. R. and Bernard, B. P. (1996). Epidemiology of work-
related musculoskeletal disorders. The orthopedic clinics of North
America, 27(4), 679-709.
62. Han, S. H. and Hong, S.W. (2003). A systematic approach for coupling user
satisfaction with product design. Ergonomics, 46 (13-14), 1441-1461.
63. Han, S. H., Hwan Yun, M., Kim, K. J. and Kwahk, J. (2000). Evaluation of
product usability: development and validation of usability dimensions and
design elements based on empirical models. International Journal of Industrial
Ergonomics, 26(4), 477-488.
64. Han, S. H., Kim, K. J., Yun, M. H., Hong, S. W. and Kim, J. (2004). Identifying
mobile phone design features critical to user satisfaction. Human Factors and
Ergonomics in Manufacturing and Service Industries, 14 (1), 15-29.
65. Harih, G. (2014).Decision support system for generating ergonomic tool-
handles. International Journals of Simulation Modeling, 13 (1), 5-15.
66. Haslegrave, C. M. (1994). What do we mean by a ‘working posture?
Ergonomics, 37(4), 781-799.
67. Helander, M. G. and Burri, G. J. (1995). Cost effectiveness of ergonomics and
quality improvements in electronics manufacturing. International Journal of
68. Hilton, M. F. and Whiteford, H. A. (2010). Associations between
psychological distress, workplace accidents, workplace failures and workplace
successes. International Archives of Occupational and Environmental
Health, 83(8), 923-933.
69. Hingtgen, B. A., McGuire, J. R., Wang, M., & Harris, G. F. (2004, September).
Quantification of reaching during stroke rehabilitation using unique
upper extremity kinematic model. In Engineering in Medicine and
Biology Society, 2004. IEMBS'04. 26th Annual International Conference of the
IEEE, 2, 4916-4919.
126
70. Hobson, D. A. (1992). Comparative effects of posture on pressure and shear at
the body-seat interface. Journal of Rehabilitation Research and Development,
29(4) 21-31.
71. Horiguchi, A. and Suetomi, T. (1995). A Kansei engineering approach to a
driver/vehicle system. International Journal of Industrial Ergonomics, 15(1), 25-
37.
72. Hong, S. W., Kim, K. J. and Han, S. H. (2002, September). Optimal balancing
of product design features: A case study on mobile phones. Proceedings of the
Human Factors and Ergonomics Society 46th Annual Meeting, 46(6), 700-704.
SAGE Publications.
73. Hua, S. K., Lu, M. T. and Tzeng, G. H. (2014). Exploring smart phone
improvements based on a hybrid MCDM model. Expert Systems with
Applications, 41(9), 4401-4413.
74. Hu, B., Ma, L., Zhang, W., Salvendy, G., Chablat D. and Bennis F. (2011).
Predicting real-world ergonomic measurements by simulation in a virtual
environment. International Journal of Industrial Ergonomics, 41(1), 64-71.
75. Hunting, W., and Grandjean, T. L. E. (1981). Postural and visual loads at VDT
workplaces. I. Constrained postures. Ergonomics, 24(12), 917–931.
76. Hwang, C.L. and Lin, M.J. (1987).Group Decision Making under Multiple
Criteria: Methods and Applications, Springer, Berlin.
77. Isiklar G. and Buyukozkan G. (2007).Using a multi- criteria decision making
approach to evaluate mobile phone alternatives, Computer Standards and
Interfaces, 29(2),265-274.
78. Jahan, A., Ismail, M. Y., Mustapha, F. and Sapuan, S. M. (2010). Material
selection based on ordinal data. Materials and Design, 31(7), 3180-3187.
79. Jang, J.R. (1991). Fuzzy modeling using generalized neural networks and
Kalmman filter algorithm. Proceedings of Ninth National Conference on Artificial
Intelligence, 762–767.
80. Jang, J.R.S. (1993). ANFIS: adaptive-network-based fuzzy inference
system. Systems, Man and Cybernetics. IEEE Transactions on Systems, 23(3),
665–685.
81. Jayaram, U., Jayaram, S., Shaikh, I., Kim, Y. and Palmer, C. (2006).
Introducing quantitative analysis methods into virtual environments for real-time
and continuous ergonomic evaluations. Computers in Industry, 57(3), 283–296.
127
82. Jee, D. H. and Kang, K. J. (2000). A method for optimal material selection aided
with decision making theory. Materials and Design, 21(3), 199-206.
83. Jeong, B.Y. and Park, K.S. (1990). Sex differences in anthropometry for school
furniture design. Ergonomics, 33 (12), 1511–1521.
84. Jeya Girubha, R. and Vinodh, S. (2012). Application of fuzzy VIKOR and
environmental impact analysis for material selection of an automotive
component. Materials and Design, 37, 478-486.
85. Jiang, H. M., Kwong, C.K., Lp, W.H. and Wong, T.C. (2011). Modeling
customer satisfaction for new product development using a PSO based ANFIS
approach. Applied Soft Computing, 12(2), 726–734.
86. Jindo, T., and Hirasago, K. (1997). Application studies to car interior of Kansei
engineering. International Journal of Industrial Ergonomics, 19(2), 105–114.
87. Jindo, T., Hirasago, K. and Nagamachi, M. (1995). Development of a design
support system for office chair using 3-D graphics. International Journal of
Industrial Ergonomics, 15(1), 49–62.
88. Jordan, P.W. (1998). Human factors for pleasure in product use. Applied
Ergonomics, 29 (1), 25-33.
89. Jung, H.S. (2005). A prototype of an adjustable table and an adjustable chair
for schools. International Journal of Industrial Ergonomics, 35(10), 955–969.
90. Jung, E. S., Kee, D. and Chung, M. K. (1992, October). Reach posture
prediction of upper limb for ergonomic workspace evaluation. In Proceedings of
the Human Factors and Ergonomics Society Annual Meeting, 36(10), 702-706.
SAGE Publications.
91. Jung, E. S., Kee, D. and Chung, M. K. (1995). Upper body reaches posture
prediction for ergonomic evaluation models. International Journal of Industrial
Ergonomics, 16(2), 95-107.
92. Jung, E. S., Shin, Y. and Kee, D. (2000).Generation of visual fields for
ergonomic design and evaluation. International Journal of Industrial
Ergonomics, 26(4), 445-456.
93. Khanam, C.N., Reddy, M.V., Mrunalini, A., 2006. Opinion of students on seating
furniture used in class room. Journal of Human Ecology, 20(1), 15–20.
94. Kaur M., Mahajan, A. Singh, K. B. Kaur A. A. and Batra, A.P.S. (2011).
Estimation of stature from upper arm length in north Indians - an anthropometric
128
study, Indian Journal of Fundamental and Applied Life Sciences, 1 (4),
151-154.
95. Kazmierczak, K., Neumann, W. P. and Winkel, J. (2007). A case study of
serial‐flow car disassembly: Ergonomics, productivity and potential system
performance. Human Factors and Ergonomics in Manufacturing & Service
Industries, 17(4), 331-351.
96. Kee, D. (2002). A method for analytically generating three-dimensional
isocomfort workspace based on perceived discomfort. Applied Ergonomics,
33(1), 51-62.
97. Kee, D., Jung, E. S. and Chung, M. K. (1992). Iso response time regions for the
evaluation of visual search performance in ergonomic interface models.
Ergonomics, 35(3), 243-252.
98. Kingma, I. and van Dieën, J. H. (2009). Static and dynamic postural loadings
during computer work in females: Sitting on an office chair versus sitting on an
exercise ball. Applied Ergonomics, 40(2), 199-205.
99. Kosiak, M., Kubicek, W.G., Wilson, M., Danz, J.N., Kottke, F.J., 1958.
Evaluation of pressure as a factor in the production of ischial ulcers. Archives of
Physical Medicine and Rehabilitation, 39 (10), 623-629.
100. Kumar, A., Bhaskar, G. and Singh, J. K. (2009). Assessment of controls layout
of Indian tractors. Applied Ergonomics, 40(1), 91-102.
101. Kuoppala, J., Lamminpaa, A. and Husman, P. (2008). Work health
promotion, job well-being, and sickness absences-a systematic review and
meta-analysis. Journal of Occupational and Environmental Medicine, 50(11),
1216-1227.
102. Kwong, C. K., Wong, T. C., & Chan, K. Y. (2009). A methodology of generating
customer satisfaction models for new product development using a neuro-fuzzy
approach. Expert Systems with Applications, 36(8), 11262-11270.
103. Lili, W., Tao, S., Zheng, Y. and Yixiong, L. (2010, June). The parametric
design of office chairs based on ergonomics. In Computing, Control and
Industrial Engineering (CCIE), 2010 International Conference, 1, 355-360.
IEEE.
104. Lindegård, A., Karlberg, C., Wigaeus Tornqvist, E., Toomingas, A. and
Hagberg, M. (2005). Concordance between VDU-users’ ratings of comfort and
129
perceived exertion with experts’ observations of workplace layout and working
postures. Applied Ergonomics, 36(3), 319-325.
105. Linder-Ganz, E., Engelberg, S., Scheinowitz, M. and Gefen, A. (2006).
Pressure–time cell death threshold for albino rat skeletal muscles as related to
pressure sore biomechanics. Journal of Biomechanics, 39(14), 2725-2732.
106. Linder-Ganz, E. and Gefen, A. (2009). Stress analyses coupled with damage
laws to determine biomechanical risk factors for deep tissue injury during
sitting. Journal of Biomechanical Engineering, 131(1), 011003(1-13).
107. Linder-Ganz, E., Shabshin, N., Itzchak, Y. and Gefen, A. (2007). Assessment of
mechanical conditions in sub-dermal tissues during sitting: a combined
experimental-MRI and finite element approach. Journal of Biomechanics, 40(7),
1443-1456.
108. Lin, M. C., Wang, C. C., Chen, M. S. and Chang, C. A. (2008). Using AHP and
TOPSIS approaches in customer-driven product design process. Computers in
Industry, 59(1), 17-31.
109. Lin, R. and Kang, Y.Y. (2000). Ergonomic design of desk and chair for primary
school students in Taiwan. Proceedings of the 4th International Conference on
Universal Access in Human Computer Interaction: Coping with Diversity,
207–216.
110. Lin, Y. C., Lai, H. H. and Yeh, C. H. (2007). Consumer-oriented product form
design based on fuzzy logic: A case study of mobile phones. International
Journal of Industrial Ergonomics, 37(6), 531-543.
111. Lim, J. and Hoffmann, E. (1997). Appreciation of the zone of convenient reach
by naive operators performing an assembly task. International Journal of
112. Lucas, R. L. (1974, October). Human Factors in Industry-Getting the Point
Across. In Proceedings of the Human Factors and Ergonomics Society Annual
Meeting, 18(4), 404-405. SAGE Publications.
113. Luczak, H., Volpert, W., Raeithel, A. and Schwier, W. (1987).
Arbeitswissenschaft–Kerndefinition. Gegenstandskatalog, Forschungsgebiete,
Eschborn: RKW, S, 59.
114. Liu, Y., Kiang, M. and Brusco, M. (2012). A unified framework for market
segmentation and its applications. Expert Systems with Applications, 39(11),
10292–10302.
130
115. Lowe, A. and Lakes, R. S. (2000). Negative Poisson's ratio foam as seat
cushion material. Cellular Polymers, 19(3), 157-168.
116. Ma, L., Zhang, W., Chablat, D., Bennis, F. and Guillaume, F. (2009). Multi-
objective optimisation method for posture prediction and analysis with
consideration of fatigue effect and its application case. Computers and
Industrial Engineering, 57(4), 1235-1246.
117. Maire, M. and Ross-Motta, J, (2000), Work-related Musculoskeletal Disorders,
Ecology research group, University of Oregon, [online]. US: Research alert.
Available: http://www.cbs.state.or.us/external/imd/rasums/resalert/msd/html
[assessed on 29 November 2012].
118. Mills, N. (2007). Polymer Foams Handbook, Engineering and Biomechanics
Applications and Design Guide. Butterworth-Heinemann.
119. Maniya, K. and Bhatt, M. G. (2010). A selection of material using a novel type
decision-making method: preference selection index method. Materials and
Design, 31(4), 1785-1789.
120. Margaritis, S. and Marmaras, N. (2007). Supporting the design of office layout
meeting ergonomics requirements. Applied Ergonomics, 38(6), 781-790.
121. Marras, W. S. and Schoenmarxlin, R. W. (1993). Wrist motions in industry.
Ergonomics, 36(4), 341-351.
122. Menendez, C.C., Amick, B.C., Robortson, M., Bazzani, L., Derango, K., Roony,
T. and Moore, A. (2012). A replicated field intervention study evaluating the
impact of a highly adjustable chair and office ergonomics training on visual
symptoms. Applied Ergonomics, 43(4), 639–644.
123. Mittal, R.K. and Nagrath, I.J. (2007). Robotics and Control. Tata McGraw-Hill,
Delhi.
124. Moes, C. C. M. N. and Horvath, I. (2002). Using finite element model of human
body for shape optimization of seats: optimization material properties.
International Design Conference, Dubrovnik. May 14-17.
125. Mokdad, M. and Al-Ansari, M. (2009). Anthropometrics for the design of
Bahraini school furniture. International Journal of Industrial Ergonomics, 39(5),
728-735.
126. Mokhlis, S. and Yaakop, A. Y. (2012). Consumer choice criteria in mobile
phone selection: An investigation of malaysian university students. International
Review of Social Sciences & Humanities, 2(2), 203-212.
131
127. Thariq, M. G., Munasinghe, H. P. and Abeysekara, J. D. (2010). Designing
chairs with mounted desktop for university students: Ergonomics and
comfort. International Journal of Industrial Ergonomics, 40(1), 8-18.
128. Mohanty, P. P. and Mahapatra, S. S. (2014). An integrated approach for
designing office chair with ergonomic consideration. International Journal of
Services and Operations Management, 17(2), 194-220.
129. Morag, I. (2007). Intel's incident-free culture: A case study. Applied
Ergonomics, 38(2), 201-211.
130. Moreau, M. (2003). Corporate ergonomics programme at automobiles Peugeot-
Sochaux. Applied Ergonomics, 34(1), 29-34.
131. Myers, R. H. (2000). Classical and modern regression with applications
(Duxbury Classic). Duxbury Press, Pacific Grove.
132. Murray, I. A. and Johnson, G. R. (2004). A study of the external forces and
moments at the shoulder and elbow while performing everyday tasks. Clinical
Biomechanics, 19(6), 586-594.
133. Musa, A.I. (2011). Anthropometric evaluation and assessment of school
furniture design in Nigeria: a case study of secondary schools in rural area of
Odida, Nigeria. International Journal of Industrial Engineering Computations,
2(3), 499–508.
134. Nagamachi, M. (1995). Kansei engineering: a new ergonomic consumer-
oriented technology for product development. International Journal of Industrial
Ergonomics, 15(1), 3-11.
135. Nagamachi, M. (2002). Kansei engineering as a powerful consumer-oriented
technology for product development. Applied Ergonomics, 33(3), 289-294.
136. Neumann, W. P., Winkel, J., Medbo, L., Magneberg, R. and Mathiassen, S. E.
(2006). Production system design elements influencing productivity and
ergonomics: A case study of parallel and serial flow strategies. International
Journal of Operations & Production Management, 26(8), 904-923.
137. Noruzy, A., Hayat, A. A., Rezazadeh, A., Najafi, S. and Hatami-Shirkouhi, L.
(2011). Factors influencing the productivity of knowledge workers: a case study
from an Iranian Oil Company. International Journal of Productivity and Quality
Management, 8(4), 459-479.
132
138. Nunes, I. L. (2009). FAST ERGO_X – a tool for ergonomic auditing and work-
related musculoskeletal disorders prevention. WORK: A Journal of Prevention,
Assessment, & Rehabilitation, 34(2), 133-148.
139. Odgen, R.W. (1997). Non-linear Elastic Deformations. Dover Publications,
New York.
140. Opricovic, S. and Tzeng, G. H. (2002). Multi criteria planning of
post‐earthquake sustainable reconstruction. Computer‐Aided Civil and
Infrastructure Engineering, 17(3), 211-220.
141. Opricovic, S. and Tzeng, G. H. (2004). Compromise solution by MCDM
methods: A comparative analysis of VIKOR and TOPSIS. European Journal of
Operational Research, 156(2), 445-455.
142. Opricovic, S. and Tzeng, G. H. (2007). Extended VIKOR method in comparison
with outranking methods. European Journal of Operational Research, 178(2),
514-529.
143. Pennestrì, E., Valentini, P. P. and Vita, L. (2005). Comfort analysis of car
occupants: comparison between multibody and finite element models.
International Journal of Vehicle Systems Modeling and Testing, 1(1), 68-78.
144. Peterson, N.C., 1976. The development of pressure sores during
hospitalization. In: Kenedi, R.M., Cowden, J.M., Scales, J.T. (Eds.), Bedsore
Biomechanics. Macmillan, London, 219-224.
145. Pressure ulcers prevalence, cost and risk assessment: consensus development
conference statement e the National Pressure Ulcer Advisory Panel. Decubitus
2, May 2, 1989, 24-28.
146. Park, J. and Han, S. H. (2004). A fuzzy rule-based approach to modeling
affective user satisfaction towards office chair design. International Journal of
147. Park, J., Snell, W., Ha, S. and Chung, T. L. (2011). Consumers’ post-adoption
of M-services: Interest in future M-services based on consumer evaluations of
current M-services. Journal of Electronic Commerce Research, 12(3), 165-175.
148. Park, M. Y., Kim, J. Y. and Shin, J. H. (2000). Ergonomic design and evaluation
of a new VDT workstation chair with keyboard–mouse support. International
133
149. Park, T. and Kim, K. J. (1998). Determination of an optimal set of design
requirements using house of quality. Journal of Operations Management, 16(5),
569-581.
150. Peterson, N. F. (1997). Production system and individual mechanical exposure,
Lic. Eng, Lund University. Lund. ISBN ISSN. 1104-1080.
151. Punnett, L. and Wegman, D. H. (2004). Work-related musculoskeletal
disorders: the epidemiologic evidence and the debate. Journal of
Electromyography and Kinesiology, 14(1), 13-23.
152. Ragan, R., Kernozek, T. W., Bidar, M. and Matheson, J. W. (2002). Seat-
interface pressures on various thicknesses of foam wheelchair cushions: a finite
modeling approach. Archives of Physical Medicine and Rehabilitation, 83(6),
872-875.
153. Rajan, V. N., Sivasubramanian, K. and Fernandez, J. E. (1999). Accessibility
and ergonomic analysis of assembly product and jig designs. International
154. Rao, R. (2012). A note on “An alternative multiple attribute decision making
methodology for solving optimal facility layout design selection problems”.
International Journal of Industrial Engineering Computations, 3(3), 519-524.
155. Rao, R. V. and Patel, B. K. (2010). A subjective and objective integrated
multiple attribute decision making method for material selection. Materials and
Design, 31(10), 4738-4747.
156. Rao, R. V. and Rajesh, T. S. (2009). Software Selection in Manufacturing
Industries Using a Fuzzy Multiple Criteria Decision Making Method,
PROMETHEE. Intelligent Information Management, 1(3), 159-165.
157. Ray, G. G., Ghosh, S. and Atreya, V. (1995). An anthropometric survey of
Indian schoolchildren aged 3–5 years. Applied Ergonomics, 26(1), 67-72.
158. Reitenbach, E.F.R., Jochems, A.J., Molenbrock, J.F.M., Ball, R. and Eijk, D.J.V.
(2009). User experience of office chairs and anthropometric of female Chinese
and Hong Kong Chinese office and factory workers. The Ergonomics Open
Journal, 2, 1–12.
159. Resnick, M. L. and Zanotti, A. (1997). Using ergonomics to target productivity
improvements. Computers and Industrial Engineering, 33(1), 185-188.
160. Suykens, J. A. and Vandewalle, J. (1999). Least squares support vector
machine classifiers. Neural Processing Letters, 9(3), 293-300.
134
161. Saaty, T. L. (1977). A scaling method for priorities in hierarchical structures.
Journal of Mathematical Psychology, 15(3), 234-281.
162. Schmidtke, H., 1989. Arbeitsplatze und arbeitsraüme in werkstatt und betrieb.
In: Brokmann, W. (Ed.), Arbeitsgestaltung in Production und Verwaltung.
Wirtschaftsverlag Bachem, Cologne.
163. Schnauber, H. (1986). Ergonomics and productivity as reflected in a new
factory. In Trends in ergonomics/human factors III, Karwowski Ed., Elsevier
Science Publishers, 459-465.
164. Schrodt, M., Benderoth, G., Kühhorn, A. and Silber, G. (2005). Hyperelastic
description of polymer soft foams at finite deformations. Technische Mechanik,
25(3-4), 162-173.
165. Shabshin, N., Zoizner, G., Herman, A., Ougortsin, V. and Gefen, A. (2010). Use
of weight bearing MRI for evaluating wheel chair cushions based on internal
soft tissue deformations under ischial tuberocities. Journal of Rehabilitation
Research and Development, 47 (1), 31-42.
166. Shanian, A. and Savadogo, O. (2006). TOPSIS multiple-criteria decision
support analysis for material selection of metallic bipolar plates for polymer
electrolyte fuel cell. Journal of Power Sources, 159(2), 1095-1104.
167. Shimizu, Y. and Jindo, T. (1995). A fuzzy logic analysis method for evaluating
human sensitivities. International Journal of Industrial Ergonomics, 15(1), 39-
47.
168. Siefert, A., Pankoke, S., and Wölfel, H. P. (2008). Virtual optimization of car
passenger seats: Simulation of static and dynamic effects on drivers’ seating
comfort. International Journal of Industrial Ergonomics, 38(5), 410-424.
169. Shackel, B. (1984). The concept of usability. Visual display terminals: Usability
issues and health concerns, Prentice-Hall, Englewood Cliffs, NJ, 45-87.
170. Shikdar, A. A. and Das, B. (1995). A field study of worker productivity
improvements. Applied Ergonomics, 26(1), 21-27.
171. Silver-Thron, M. B. and Childress, D.S. (1996). Parameter analysis using finite
element method to investigate prosthetic interface stresses for persons with
trans-tibial ampulation. Journal of Rehabilitation Research and Development,
33(3), 227-238.
135
172. Singh, B., Kaur, M., Kaur, J., Singh, M. and Batra, A.P.S., (2013). Estimation of
stature from forearm length in north indians - an anthropometric study.
International Journal of Basic and Applied Medical Sciences, 3 (1), 201-204.
173. Sobeih, T., Salem, O., Genaidy, A., Abdelhamid, T. and Shell, R. (2009).
Psychosocial factors and musculoskeletal disorders in the construction
industry. Journal of Construction Engineering and Management, 135(4), 267-
277.
174. Sugeno, M., and Kang, G. T. (1988). Structure identification of fuzzy model.
Fuzzy Sets and Systems, 28(1), 15-33.
175. Takagi, T. and Sugeno, M. (1985) .Fuzzy identification of systems and its
applications to modeling and control. Systems, Man and Cybernetics, 15(1),
116–132.
176. Tang, C.Y., Chan, W. and Tsui, C.P., (2010). Finite element analysis of contact
pressures between seat cushion and human buttock-thigh tissue, Journal of
Engineering, 2(9), 720-726.
177. Tang, C.Y. and Tsui, C.P. (2006). Method of Modeling Muscular Tissue with
Active Finite Elements. U. S. Patent, the Hong Kong Polytechnic University,
Kowloon (HK).
178. Tewari, V. K. and Prasad, N. (2000). Optimum seat pan and back-rest
parameters for a comfortable tractor seat. Ergonomics, 43(2), 167-186.
179. Thakurta,K.,Koester,D.,Bush,N.,Bachle,S.,1995.Evaluating short and long term
seating comfort,in:SAE Technical paper series 950144(Troy,Mi:Society of
Automotive Engineers),33-37.
180. Todd, B. A. and Thacker, J. G. (1994). Three dimensional computer models of
the human buttocks, in Vivo, Journal of Rehabilitation Research and
Development, 31(2), 111-119.
181. Tzeng, G. H., Lin, C. W. and Opricovic, S. (2005). Multi-criteria analysis of
alternative-fuel buses for public transportation. Energy Policy, 33(11), 1373-
1383.
182. Tzeng, G. H., Tsaur, S. H., Laiw, Y. D. and Opricovic, S. (2002). Multi criteria
analysis of environmental quality in Taipei: public preferences and improvement
strategies. Journal of Environmental Management, 65(2), 109-120.
183. Vergara, M. and Page, A. (2000). System to measure the use of the backrest in
sitting-posture office tasks. Applied Ergonomics, 31(3), 247-254.
136
184. Verver, M. M., Van Hoof, J., Oomens, C. W. J., Wismans, J. S. H. M. and
Baaijens, F. P. T. (2004). A finite element model of the human buttocks for
prediction of seat pressure distributions. Computer Methods in Biomechanics
and Biomedical Engineering, 7(4), 193-203.
185. Verver, M. M., De Lange, R., van Hoof, J. and Wismans, J. S.H.M. (2005).
Aspects of seat modelling for seating comfort analysis. Applied
Ergonomics, 36(1), 33-42.
186. Vignais, N. and Marin, F. (2014). Analysis of the musculoskeletal system of the
hand and forearm during a cylinder grasping task. International Journal of
187. Vink, P., Koningsveld, E. A. and Molenbroek, J. F. (2006). Positive outcomes of
participatory ergonomics in terms of greater comfort and higher productivity.
188. Veryzer, R. W. (1995). The place of product design and aesthetics in consumer
research. Advances in Consumer Research, 22, 641-645.
189. Vos, G. A., Congleton, J. J., Steven Moore, J., Amendola, A. A. and Ringer, L.
(2006). Postural versus chair design impacts upon interface pressure. Applied
Ergonomics, 37(5), 619-628.
190. Wang, K. C. (2011). User-oriented product form design evaluation using
integrated Kansei engineering scheme. Journal of Convergence Information
Technology, 6(6), 420-438.
191. Wang, X. (1999). A behavior-based inverse kinematics algorithm to predict arm
prehension postures for computer-aided ergonomic evaluation. Journal of
Biomechanics, 32(5), 453-460.
192. Wang, Y. C. and Lakes, R. (2002). Analytical parametric analysis of the contact
problem of human buttocks and negative Poisson's ratio foam cushions.
International Journal of Solids and Structures, 39(18), 4825-4838.
193. Wang, M. J. J., Wang, E. and Lin, Y. C. (2002). The anthropometric database
for children and young adults in Taiwan. Applied Ergonomics, 33(6), 583-585.
194. Wang, X. and Verriest, J. P. (1998). A geometric algorithm to predict the arm
reaches posture for computer-aided ergonomic evaluation. The Journal of
Visualization and Computer Animation, 9(1), 33-47.
137
195. Weiman, N. (1982, October). Integrating Human Factors into Product
Development within Industry. In Proceedings of the Human Factors and
Ergonomics Society Annual Meeting, 26(8), 731-732. SAGE Publications.
196. Westgaard, R. H. and Aarås, A. (1984). Postural muscle strain as a causal
factor in the development of musculo-skeletal illnesses. Applied Ergonomics,
15(3), 162-174.
197. Wu, C.F.J., Hamada, M., 2000.Experiments: Planning, Analysis and Parameter
Design Optimization, Wiley-India, New Delhi.
198. Yang, S. M., Nagamachi, M. and Lee, S. Y. (1999). Rule-based inference
model for the Kansei Engineering System. International Journal of Industrial
Ergonomics, 24(5), 459-471.
199. Yoon, K. P. and Hwang, C. L. (1995). Multiple attribute decision making: an
introduction, 104, Sage Publications.
200. Yu, P. L. (1973). A class of solutions for group decision problems. Management
Science, 19(8), 936-946.
201. Yue ZL (2011). Deriving decision maker’s weights based on distance measure
for interval-valued intuitionistic fuzzy group decision making. Expert System
Application, 38(9), 11665-11670.
202. Zhao, J. and Badler, N. I. (1994). Inverse kinematics positioning using nonlinear
programming for highly articulated figures. ACM Transactions on Graphics
(TOG), 13(4), 313-336.
203. Zadeh, L. A. (1975).The concept of a linguistic variable and its application to
approximate reasoning. Information Sciences, 8 (3), 199–249.
138
APPENDIX
List of Publications
International Journals (Published)
1. Mohanty, P. P., Mahapatra, S. S. (2014). A finite element approach for analyzing the
effect of cushion Type and thickness on pressure ulcer. International Journal of
2. Mohanty, P. P., Mahapatra, S. S. (2014). An integrated approach for designing office
chair with ergonomic consideration. International Journal of Services and Operations
Management. 17(2), 194-220.
3. Mohanty, P. P., Mahapatra, S. S. (2013). Design of office chair: A quality function
deployment approach. Advanced Materials Manufacturing and Characterization. 3(2),
520-523.
International Journal (Communicated)

1. Mohanty, P.P., Mahapatra, S. S. (2014). A novel multi-attribute decision making
approach for selection of appropriate product conforming ergonomic considerations.
Behavior and Information Technology (Under review).
International Conferences
1. Mohanty, P. P., Mahapatra, S. S. (2013). Seating comfort in office environment:

numerical analysis on human soft tissue and seat cushion. Third International
Conference on Production and Industrial Engineering (CPIE 2013) held on 29-31
March at National Institute of Technology, Jalandhar.
2. Mohanty, P. P., Mahapatra, S. S. (2013). A finite element analysis for stress
distribution in muscle tissue of buttock for different cushion thickness. Second
International Conference on Industrial Engineering (ICIE 2013) held on 20-22
November 2013 at National Institute of Technology, Surat.
3. Mohanty, P. P., Mahapatra, S. S. (2014). Application of MADM method for selecting
ergonomically designed office chair. 1st International Conference on Mechanical
a
Engineering: Emerging Trends for Sustainability (ICMEETS-2014) held on 29th-31st
January 2014 at National Institute of Technology, Bhopal.
4. Mohanty, P. P., Mahapatra, S. S. (2014). A compromise solution by VIKOR method
for ergonomically designed product with optimal set of design characteristics”,
Second Annual International Conference on Material proceedings and
Characterization held on 8th-9th March 2014 at GRIET, Hyderabad.

Studies On Product Design Using Ergonomic Considerations: Doctor of Philosophy

Uploaded by

Copyright:

Available Formats

Studies On Product Design Using Ergonomic Considerations: Doctor of Philosophy

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Studies On Product Design Using Ergonomic Considerations: Doctor of Philosophy

Uploaded by

Copyright:

Available Formats

Studies on Product Design using

A THESIS SUBMITTED IN FULFILLMENT OF

THE REQUIREMENT FOR THE AWARD OF THE DEGREE

DEPARTMENT OF INDUSTRIAL DESIGN

PRAGYAN PARAMITA MOHANTY

NATIONAL INSTITUTE OF TECHNOLOGY

This is to certify that the thesis entitled STUDIES ON PRODUCT DESIGN

USING ERGONOMIC CONSIDERATIONS submitted by Pragyan Paramita

Mohanty has been carried out under my supervision in fulfillment of the

requirement for the award of the degree of Doctor of Philosophy in

Industrial Design at National Institute of Technology Rourkela and this

Prof. Siba Sankar Mahapatra

Department of Industrial Design

PRAGYAN PARAMITA MOHANTY

1 Background and Motivation

3 An Integrated Approach for Designing Office Furniture

4 A Numerical Approach for Ergonomic Design

5 A Novel Multi-Attribute Decision Making Approach for

6 Human Upper Arm Posture Prediction within

7 Executive Summary and Conclusions

5.18 Ranking of alternatives considering integrated weight 95

Table Caption Page

These problems offer new opportunities for ergonomic design as follows:

elastic modulus of E=200kPa and density  =60kg/m3. The model predicts a

Layout design for

Figure 2.1 Percentage of papers under each classification category

2.2 Ergonomic consideration for improving user satisfaction

2.3 Bio-mechanical analysis in ergonomic design

2.4 Ergonomic consideration in product selection

2.5 Layout design for improving work environment

Ergonomically designed industrial or office work environment consider both

3.2.1 Data collection

8 The chair is made for active sitting 1 2 3 4 5

25 The chair has a personal recognition 1 2 3 4 5

Table 3.2 Tangible design attributes

Table 3.3 Intangible design attributes

Table 3.4 Anthropometric standing dimension (mm)

3.2.2 Factor Analysis

customer rating, A ij , denote the relative importance of i th characteristic with respect to

3.2.4 Adaptive Neuro Fuzzy Inference System (ANFIS):

For convenience outputs of this layer will be called normalized firing

Thus, an adaptive network is presented in Figure 3.1 is functionally equivalent to a

Figure 3.1 A typical architecture of ANFIS structure

3.3 Results and discussions

Initial rating of customer requirements for each model is derived using a 1 to 10

No Design requirement Initial Design Revised Normalized Rank

Table 3.9 Ranking of Design attributes for model-2(balance)

No QFD model 2(balance) Initial Revised Normalized Rank

Table 3.10 Ranking of Design attributes for model-3(luxuriousness)

No Design requirement Initial Design Revised Normalized Rank

9 Shape of seat pan 4.482 6.696 0.652 8

Figure 3.11 Effect of control design parameters on response

Sl. No Design attributes Optimal values BIS

The stress distribution throughout meshed model is studied by varying the

parameters to be determined, J el and J th are the elastic volumetric deformation and

thermal volumetric deformation respectively, ˆ 1 , ˆ 2 , ˆ 3 are the principal stretch ratios

which provide a measure of deformation.

where  i (relaxation time),Gi (relaxation modulus) and G0 (instantaneous shear

Table 4.2 Coefficients of Prony series parameters for cushion exhibiting