Evaluation of physical fatigue based on motion analysis in manual handling of loads Bernardo et al.

Evaluation of physical fatigue based on motion analysis in manual

handling of loads

F. G. Bernardoa, R. P. Martinsb, J. C. C. Guedesc

a
Faculty of Engineering, University of Porto, PT ([email protected]) ORCID: 0000-0002-8935-5689, bLAETA,
Faculty of Engineering, University of Porto, PT ([email protected]) ORCID: 0000-0001-8438-0597, C LAETA, Faculty of
Engineering, University of Porto, PT ([email protected]) ORCID: 0000-0003-2367-2187.

Article History Abstract

Received 27 June 2018 Muscle fatigue refers to the transient decrease in the capacity to
Accepted 25 October 2018 perform physical actions and can cause productivity loss, human
Published 5 April 2019 errors, unsafe actions, injuries and work-related musculoskeletal
disorders (WMSDs). A total of 13 participants repetitively lifted a
Keywords 2.5 kg load at a total elevation of 0.5 m, until voluntary exhaustion
or intensive pain. Several indicators of muscle fatigue were found,
Motion analysis
including increased forward bending, micro-expressions, changing
Manual handling of loads
load support strategy, holding load closer to the chest during
Movement variability
elevation and increase in eccentric movement speed. Volunteers
Repetitive movements
that practiced sports regularly lasted longer in the experiment and it
Biomechanics
was found that smoking and sedentarism limited the exercise
capacity of some subjects. Volunteers increased the wrists and
DOI: elbows velocity during the experiment. It was verified that half of
10.24840/2184-0954_003.001_0002 the volunteers had less fluid movement during lifting, with higher
ISSN: acceleration values in some parts and lower values in others,
2184-0954 translating into more impulsive movements. It was concluded that
to compensate muscle fatigue, people adapt their working strategy,
Type: changing movement patterns, recruiting different muscles and
Research Article changing kinetic or kinematic components of the movement (like
Open Access joint angles and velocities).
Peer Reviewed
CC BY

1. INTRODUCTION

According to Enoka et al. (2008), muscle fatigue refers to the transient decrease in the
capacity to perform physical actions and the decrease in the maximal force or power (velocity
of muscle contraction) that the involved muscles can produce. It can be categorized as one of
the symptoms of blood occlusion, because limited blood flow delivers insufficient oxygen and
nutrients, alongside the inadequate removal of metabolic waste products, causing lactate
concentrations to rise (Oyewole, 2014).
Manual handling of loads consist of any of the following activities: lifting, holding, putting
down, pushing, pulling, carrying or moving of a load. Fatigue accumulation can be extended by
practicing proper work/rest time ratio and controlling the lifting variables (Halim et al., 2014).
A significant level of muscle fatigue can cause productivity loss, human errors, unsafe actions,
injuries and work-related musculoskeletal disorders (WMSDs) (Sluiter et al., 2003; Toole,
2005; Huang and Hinze, 2006; Hallowell, 2010).
During repetitive and fatiguing work, the musculoskeletal system adapts and uses momentary
muscle substitution patterns, which result in more variable and less coordinated movements
(Mehta et al., 2015).
A recent review, performed by Srinivasan and Mathiassen (2012) concluded that motor
variability is a relevant issue in an occupational context and that there is a great need for
International Journal of Occupational and Environmental Safety, 3:1 (2019) 13-26
Evaluation of physical fatigue based on motion analysis in manual handling of loads Bernardo et al.

studies
13

International Journal of Occupational and Environmental Safety, 3:1 (2019) 13-26

 of motor variability. They suggest future research in creating methods to assess motor
variability and study the relationship with occupational tasks and outcomes like fatigue and
performance.
The main goals of this study were:
a) Identification of muscle fatigue indicators that could be listed and checked in a real work
environment;
b) Definition of a muscle fatigue assessment method based on motion analysis, in order to
objectify a measurement technique susceptible of being applied in a real work environment.

2. MATERIAL AND METHODS

2.1. Participant selection

This methodology is a new approach that allows the assessment of common symptoms and
behaviors that people demonstrate when starting to feel muscle fatigue
Participants must be healthy, because any diseases or injuries may influence muscle
performance and lead to inconsistent findings. Variables like lifestyle, eating habits and medical
history are important.
Smoking and alcohol consumption also influence the results, alongside age and gender
differences (Al-Mulla et al., 2011). According to Wüst et al. (2008) smokers have lower skeletal
muscle fatigue resistance and the increased fatigability is due to the reduction in oxygen
supply to the muscle. It was concluded by Bogdanis (2012) that sedentary lifestyle and/or
cardiovascular and pulmonary diseases may limit exercise capacity and increase fatigability.
In order to analyze these variables, a questionnaire divided into 4 topics (personal data,
lifestyle, eating habits and medical history) was created.
The experiments were conducted in the morning between 8 a.m. and 12 p.m., in order to have
rested volunteers. Otherwise, muscle condition would be influenced by daily activities.

2.2. Sample characteristics

A total of 13 participants were recruited, namely, 4 men and 9 women and it was a
convenience sample gathered in the university where the study was performed. The analyzed
sample of 13 people is not the ideal sample size to generalize the results and future work
should include a larger and representative sample.
All volunteers were university students and had normal weight, except one of the males that
had more muscular volume than the others, which led to a higher Body Mass Index value
(BMI) for that individual. Their characteristics are presented in Table 1. All participants
provided informed written consent.

Table 1. Subjects Characteristics

Age Height (m) Weight (kg) BMI (kg/m2)

Males 24 ± 0.8 1.83 ± 0.08 76.8 ± 7.3 22.9 ± 2.2
Females 24 ± 1.2 1.65 ± 0.05 62.7 ± 3.8 23.0 ± 1.6

2.3. Apparatus/Equipment

The software used consisted of a motion analysis software available from Video4Coach, namely
SkillCapture version 2.0.6 and SkillSpector Version 1.2.4. The first one was used to record the
experiment, alongside two Logitech C920 HD Pro Webcams, each one placed in the left and
right side of the person, at around 45 degrees. The recordings were then studied in
SkillSpector.
A digitizing model and a calibration model were created. The digitizing model included a total of
12 points, namely two on the fingers, two on the wrists, two on the elbows, two on the
International Journal of Occupational and Environmental Safety, 3:1 (2019) 13-26
1
 shoulders, one on the neck and one on the top of the head. Segments were defined as an
element going

International Journal of Occupational and Environmental Safety, 3:1 (2019) 13-26

1
from one point to another. For example, the forearm was the segment between the wrist and
elbow points. The calibration model included 8 points, each in every corner of the 50 cm edge
cubic structure.
Volunteers had markers placed in strategic parts of the upper body, as shown in Figure 1, in
order to improve the video analysis and creation of the body segments in the 3D model, as
presented in Figure 2.
The 3D model was built by marking each reference point placed on the volunteer body in each
frame (1 frame = 0.03 s) of the video. Since two cameras were used, this had to be done
twice, one for each recorded video. Because of this limitation, only the differences between the
first and last movements of lifting and lowering were studied, considering that the last
movement before voluntary exhaustion is the most representative of muscle fatigue.
The body position was calculated using a DLT algorithm (Direct Linear Transformation) which
was found based on the calibration object. This is the mathematical methods used to transform
image data into real-world coordinates.

Figure 1. Body Segments and markers

Figure 2. 3D model created in SkillSpector

2.4. Experimental procedure

The study started with the completion of a short questionnaire that covered 4 topics, namely:
personal data, lifestyle, eating habits and medical history. After completing the questionnaire
and verifying that subjects were eligible for the study, the experiment was explained to the
volunteers and informed consent was signed.
The experiment was conducted inside a climatic chamber, where specific conditions were set,
namely 25 ± 2 ºC temperature and 49 ± 5 % relative humidity.
At the beginning of the study, a cubic calibration structure with 50 cm edge was placed where
the volunteers would perform the exercise, as shown in Figure 3, with the 3 adopted axis that
was used. This axis orientation was selected because the program calibration tutorial adopted
this orientation.
The volunteers repetitively lifted a 2.5 kg load (dumbbell disk) from a 0.72 m height to 1.22 m,
meaning a total elevation of 0.5 m, until voluntary exhaustion or intensive pain, that is, until
they no longer could perform the task. While performing the exercise, two video cameras
recorded the movements performed and these images were subsequently analyzed. A
metronome (100 bpm) marked the pace that volunteers needed to perform the lifting and
lowering movements of the load so that all volunteers performed the task under equal
conditions. Figure 4 shows the experiment layout and Figure 5 represents the lifting and
lowering movement executed by the participants.
Ratings of perceived exertion (RPE) were gathered using the modified Borg scale. Subjects
rated their level of fatigue on a scale from 0 (none at all) to 10 (maximal exertion) at the end
of the experiment. Subjects were also asked about the physical symptoms they experienced
throughout the experiment and these were noted down.

Figure 3. Calibration Object

Figure 4. Experiment layout

Figure 5. Lifting and lowering movement

2.5. Statistical Analysis

Normality tests were performed in order to determine the normality of the data, namely
Shapiro- Wilk, Anderson-Darling, Lillefors, and Jarque-Bera tests. Outliers were detected with
Grubbs test.
Student's paired t-test compared the means of two series of measurements performed on the
same statistical units, namely the information of the first and final movements.
Box-and-whisker plots were used to present the results.
Tests were carried out to verify the normality assumptions based on skewness and kurtosis,
considering the criteria proposed by George and Mallery (2010) of skewness and kurtosis
values within ± 2.

3. Results and Discussion

3.1. Questionnaire Results

The questionnaire covered 4 topics, namely: personal data, lifestyle, eating habits and medical
history. Table 2 sums the main results of the questionnaire.
In terms of lifestyle, 9 in 13 participants played sports regularly (3 out of 4 males and 6 out of
9 females). On average, both males and females practiced sports 3 times a week. The main
activity was training at the gymnasium.
Only 1 female smoked and 4 in 13 (2 out of 4 males and 2 out of 9 females) admitted they had
a sedentary lifestyle. In average, males sleep 8 ± 0 h/day while females sleep 7.2 ± 0.4 h/day.
 Table 2. Questionnaire results
Sedentary
Name Plays Sports Smokes Coffee Regularly
Lifestyle
Male 1 Yes No No Yes
Male 2 Yes No Yes Yes
Male 3 No No Yes Yes
Male 4 Yes No No Yes
Female 1 No Yes Yes Yes
Female 2 No No Yes No
Female 3 Yes No No Yes
Female 4 Yes No No No
Female 5 No No No Yes
Female 6 Yes No No Yes
Female 7 Yes No No No
Female 8 Yes No No Yes
Female 9 Yes No No Yes

A total of 10 people drink coffee regularly (only 3 females don’t drink). None of the participants
consume alcoholic drinks regularly or was taking a diet or was vegetarian.
Analyzing the data, there is a slight tendency for people that practice sports to hold on longer
in the experiment. Female 1 and 2 only kept doing the experiment for 7 and 10 minutes (see
Figure
6) and both of them did not practice sports regularly. In the males, male 3 was the only one
that didn’t practice it regularly but hold for 45 minutes (more 11 minutes than male 2).
The only smoker (10 cigarettes/day) was female 1, the one that only did the experiment for 7
minutes, which shows that smoking may harm physical condition. In terms of sedentarism,
female 1 and 2 are the only females that admit they have a sedentary life, showing that the
lack of exercise might have an influence on the results. The asthmatic condition of female 2
probably contributed to such a low time.
The asthmatic condition of female 2, probably contributed to such a low time. On males, the
results are not conclusive, since male 2 and 3 considered themselves sedentary and hold more
than male 1. On the other side, male 4, the person that stayed in the experiment for 1 h,
didn’t consider to have a sedentary life.
Most volunteers consumed coffee regularly, only female 2, 4 and 7 didn’t. Since they consume
it regularly, their body already adapted to a certain daily dose and side effects don’t appear.
Alterations appear when they consume a higher quantity or don’t drink it. According to (Poole
et al., 2017), coffee consumption was more often associated with benefit than harm for a
range of health outcomes. The largest relative risk reduction was obtained at intakes of three
to four cups a day versus none. Caffeine can also strengthen their ability to recover from
fatigue. In this study, the observed average consumption was 2 coffees/day.
Female 1 combined the negative effects of smoking with a sedentary life, which might explain
her lower performance.

3.2. Posture and behavioral changes recorded on camera and symptoms described by
the volunteers

According to literature, higher movement variability occurs with the development of muscle
fatigue (Brown et al., 2016). Muscular and kinematic adaptations occur to reduce the load on
the fatigued muscles (McDonald et al., 2016). Other studies concluded that fatigue-induced
changes in movement strategies (Lee et al., 2014) and subjects altered their kinematic
patterns significantly in response to muscle fatigue (Gates et al., 2011).
In order to confirm if any of these conclusions were also observed in the experiment, the
recordings were analyzed.
As the experiment went on, the first noticeable changes were micro expressions (involuntary
facial expressions), such as biting or narrowing lips during elevation and eyebrows down and
together. These were signs that volunteers were starting to feel uncomfortable. An increase in
breathing pattern also occurred, as verified by watching the recordings and also by the
volunteers’ testimony.
Volunteers started to change the way they executed the exercise by lifting and holding the
load, in different ways than the initial. The load was meant to be lifted horizontally relative to
the table. Some volunteers started to rotate the disk during the elevation and making it
vertical to the ground, at maximum elevation. Others opted to sustain it from below, with the
hands together and palms facing up. The ones that kept the hands in the intended position,
started to tighten and release the fingers on the load, trying to get a better grip and also,
enable some rest in each hand.
The movements started to be less fluid and while lowering the load, subjects increased
eccentric movement speed, using gravity to help them lower the load.
At the start of the lifting and when they inverted lifting to lowering, subjects demonstrated
increasing stooped position, as well, as slight head bending forward. When inverting lifting to
lowering, subjects struggled to maintain their position.
Another aspect noticed was that people started bending their neck sideways, to relieve tension
on the trapezius muscle and shoulders. They also rotated their shoulder from time to time.
Due to the fact that volunteers were standing still, they started decreasing their body stability
and started balancing more. They started to hold their body on one leg and then changing to
the other. The fact that some volunteers felt discomfort in the legs (numb legs), although the
experiment only directly fatigued upper limbs, shows that there might be a correlation between
upper and lower body. By standing still and frequently keeping the arms horizontally at
shoulder level, this corresponds to static muscle work, which means that the blood requirement
was higher than the supply. This follows the conclusions made by Oyewole (2014), because
limited blood flow, delivers insufficient oxygen and nutrients. This means that when studying
muscle fatigue, the whole body should be assessed together.
Subjects started noticing they were adopting a stooped position and started stretching their
body and spine from time to time, in order to correct their posture and relief tension in their
lower back.
In the final stages of the experiment, the movement started to be performed by lifting the load
up to the chest and then moving it forward, instead of the gradual elevation and move away
from the body. This strategy was used to compensate muscle fatigue and because lifting a load
closer to the body is easier to perform.
As the experiment went on, some struggled to follow the metronome rhythm and started
executing the movement in a very fast pattern, while others performed it slowly. This might
indicate that people started losing focus.
Another important aspect is the fact that there were no external influences/incentives
contributing to the subject’s activity. Some volunteers expressed that they would probably do
better if they were listening to music, or watching a video. Others suggested that more than
one person should do the experiment at the same time, so they could talk and distract
themselves. This might mean the psychologic component is also important, making mental
fatigue influence physical performance. A recent systematic review conducted by Van Cutsem
et al. (2017) assessed the effects of mental fatigue on physical performance. It concluded that
mental fatigue impairs endurance performance and is mediated by a higher than the normal
perception of effort.
Summing up, the posture and behavioral alterations observed in the recordings included:
a) Microexpressions (involuntary facial expressions) including biting or narrowing lips during
elevation, eyebrows down and together;
b) Increased breathing pattern;
c) Changing the load support strategy (hand and wrist posture);
d) Tightening and releasing fingers on the load, trying to get a better grip;
e) Increase in eccentric movement speed, following the conclusion of Brown et al. (2016).
f) Increased forward bending, also described by Mehta et al. (2014);
g) Head slightly bending forward;
h) Neck bending sideways;
i) Decrease in body stability;
j) Changing support leg (some volunteers started to feel numb legs);
k) Stretching body and spine;
l) Holding load closer to the chest during elevation in the final stages of the experiment,
resulting in a less fluid movement. This shows volunteers adapted to compensate fatigued
muscles, which was also concluded by McDonald et al. (2016), Lee et al. (2014) and Gates et
al. (2011);
m) Increased difficulty to follow metronome rhythm;

Symptoms described by the volunteers included discomfort in the wrists, brachioradialis, biceps
brachii, triceps brachii, deltoids (anterior, medial and posterior), erector spinal muscles and
gastrocnemius muscle.

3.3. Total execution time, Borg Scale, initial and final movements time
The total execution time, Borg Scale and initial and final movements time are presented in Table
3. The total time in the experiment of each participant, both males (M) and females (F), is
presented in Figure 6, while Figure 7 presents the Borg Scale values, from 0 (none at all) to 10
(maximal exertion).

Table 3. Total time, Borg Scale, initial and final movements time
ttotal (s) Borg Scale ti (s) tf (s)
Males 40.5 ± 15.8 4.3 ± 0.9 4.8 ± 0.4 5.0 ± 0.1
Females 18.3 ± 9.9 5.3 ± 1.9 5.0 ± 0.3 4.6 ± 0.7
Total Time (min)

7
46
0
4 22 2 3 3 5 0
2 7 10 10 24236 4
0
6 9
0 13 14
0
3
1
0
5
0
0
0
F 2F 1

Figure 6. Total time in the experiment

1 8
7 7
08
6 4 5. 6 3 5
Borg Scale

4 4 5 3. 5 3
2 .
3. 5
5
0 5
F 2F 1

Figure 7. Borg Scale values

Analyzing Figure 6, it seems that the males could handle more time performing the
experiment. Male 3 and 4 had the biggest times, with 45 and 60 minutes respectively. The 3rd
biggest time belongs to female 9, with 36 minutes, followed closely by male 2 with 34 minutes.
Female 1 had the lowest time, with only 7 minutes of exercise.
Subjects rated their level of fatigue on a scale from 0 (none at all) to 10 (maximal exertion)
and the scores are presented in Figure 7. The highest scores were given by female 2, 3 and 7.
Female
2 and 3 did the same time (10 minutes) and gave similar Borg scale results (8 and 7
respectively), corresponding to a really hard exercise. Female 7 gave a similar score and did
the experiment for 24 minutes. Female 9 gave the lowest score and was the female that did
the
experiment for a longer time. On the other side, female 1 only gave a value of 4, half of female
2, which shows that it might not be self-aware of her fatigue level. In a work environment, this
can lead to excessive effort and cause injuries. Male 1 and 4 gave the same score, although
male 4 lasted three times more in the exercise.

3.4. Statistical Results

The statistical results, namely the Student's paired t-test, of the wrists, elbows, and shoulders
are presented in Figure 8, Figure 10 and Figure 13, respectively. The dashed line boxes mark
the lowest p-values obtained in terms of position, velocity, and acceleration.

A. Wrists

Figure 8 analysis can be divided into the following parts:

1) Wrists range (position) on the X axis (Range of motion):

In order to test the hypothesis of forwarding bending increase and assess if participants lifted
the load in a stooped posture that they bent more over time, as concluded by Mehta et al.
(2014), the highest position values in the X-axis were analyzed.
The normality tests showed that the results followed a normal distribution because the p-
values (bilateral) were greater than the significance level (alpha = 0.05).
In the Student's paired t-test results, as shown in Figure 8, the calculated p-value (bilateral)
for both wrists was greater than the significance level alpha = 0.05 (R Wrist = 0.47 and L Wrist
= 0.95), meaning that the initial hypothesis “The difference between the means is equal to 0”
shouldn’t be rejected. This means that no significant differences were observed in the wrist
range mean in the X-axis.
The results demonstrated that the individuals increased their wrist range on the X-Axis.
Overall, the right wrist displayed more changes in terms of interquartile range, which
translated to a much more concise pattern of the volunteers, probably because only 1 person
was left handed. Right-handed people might command that hand with more intensity and
support more weight, in order to get higher stability, while their left hand is used for balance.

2) Wrists 3D Position

The highest 3D position values were analyzed because they occurred at the inversion point
between lifting and lowering. The data followed a normal distribution.
Analyzing the p-values, no significant differences were observed in the wrist 3D position means
before and after. The right wrist displayed more changes (lower p-value).

3) Wrists 3D Velocity

In order to compare the lifting and lowering stages, the velocity analysis was divided into two
sections. In both stages, the data analyzed was the highest velocity value verified.
Starting with the lifting stage, male 2 was removed in order to have a normal distribution. The
calculated p-value (bilateral) for both wrists was greater than the significance level alpha =
0.05 (R Wrist = 0.79 and L Wrist = 0.60), meaning that no significant differences were
observed in the wrist 3D velocity means before and after.
The left wrist had a more prominent alteration when compared to the right wrist (lower p-
value). This translates into a higher variability of velocity values on the non-dominant hand.
The reason might be that the right wrist is more controlled and needs to be more stable and
regular in the movements. Nonetheless, the results showed that both wrists increased their
velocity, although some people decreased it.
On the lowering stage, female 7 was removed in order to have a normal distribution. No
significant differences were observed in the wrist 3D velocity mean before and after, but the
right wrist had lower p-value.
The results showed that the lowering velocity increased. This follows the conclusions made by
Brown et al. (2016) that stated that there could be an increase in eccentric movement speed
since the muscle cannot resist the gravitational force.

4) Wrist 3D Acceleration

The p-value (bilateral) for both wrists was greater than the significance level alpha = 0.05 (R
Wrist = 0.21 and L Wrist = 0.28), meaning that no significant differences were observed in the
wrist 3D acceleration mean before and after. The right wrist demonstrated higher differences
due to a lower p-value.

Figure 8. Wrists student's paired t-test results

Figure 9 shows differences in the right and left wrists, in the initial (I) and final (F) movements.
In the right wrist, the first quartile and the median decreased, while the mean and third
quartile increased. On the other side, the left wrist only had the same pattern with the first
quartile, since it increased the median, and decreased the mean and third quartile.

2 Lifting Stage -

3D Acceleration (m/s2)
3D Accerelation
1
,
8
R Wrist IR Wrist F L
Wrist IL Wrist F
1
,
6

1
,
4

1
,
2

0
Figure 9 – Wrists lifting8,3D acceleration variation

0
Half of the volunteers had lower acceleration , variability in the final movement, with only
6
maximum and minimum accelerations differing around 1 m/s2, while the other half increased
that difference. The half that increased had less fluid movement, with higher acceleration
values in some parts and lower values in others.
In the lowering stage, Grubbs test identified female 7 as an outlier and it was removed in order
to have a normal distribution. No significant differences were observed in the wrist 3D
acceleration means before and after.
The variation in acceleration values increased on both wrists. This complements the conclusion
of the velocity results where volunteers had a higher eccentric movement speed. Comparing
the values, the mean value in the lifting stage was 1.16 m/s2 and 1.21 m/s2 in the lowering
stage, meaning that the lowering stage had higher variation, since the volunteers would drop
the load with the help of gravity, and not lower it carefully and in a sustained way.

B. Elbows

Figure 10 analysis can be divided into the following parts:

1) Elbows Z-Axis Range

All the 13 volunteers were included. The results followed a normal distribution.
No significant differences were observed in the elbow Z-Axis means before and after.
The results showed that volunteers didn’t open more the arms, but the other way around, they
closed the elbows and hold them near the body, in order to increase stability and hold the load
better.

2) Elbows 3D Velocity
In order to compare the lifting and lowering stages, the same strategy used for the wrists was
applied, that is, the velocity analysis was divided into lifting and lowering.
In the lifting stage, all subjects were included and a normal distribution was verified.
The p-value (bilateral) for the right elbow was greater than the significance level alpha = 0.05
but not the left elbow (R Elbow = 0.43 and L Elbow = 0.04), meaning that in the left elbow
there was a significant difference in the 3D velocity mean before and after.
 Figure 10. Elbows Student's paired t-test results

Observing Figure 11, that difference is noticeable. Comparing both elbows, they increased
interquartile range, third quartile, and range between maximum and minimum. While the right
elbow almost didn’t change median, on the left elbow it rose from 0.44 to 0.52 m/s. The
results indicate that volunteers performed the lifting with higher velocity in the final movement
of the experiment, especially on the left elbow. The same conclusion was made with the wrists,
also with more evidence on the left side. This means that both left wrist and elbow had higher
differences from beginning to end.
In the lowering stage female 7 was removed in order to obtain a normal distribution. No
significant differences were observed in the elbow velocity mean values.
The results demonstrated that the lowering movement was performed with higher velocity,
following the same idea of the wrists and the conclusions made by Brown et al. (2016).
 Lifting Stage - Elbows 3D
Velocity
0,8

3D Velocity (m/s)
0,7

0,6

0,5

0,4

0,3

0,2
R Elbow I R Elbow F L Elbow I L Elbow
F

Figure 11. Elbows Lifting 3D Velocity

3) Elbows 3D Acceleration

In order to compare the lifting and lowering stages, the acceleration analysis was divided into 2
sections.
Analyzing the p-value (bilateral) for both elbows, the left elbow had the alpha value, meaning
that the differences were significant.
Observing Figure 12, both elbows increased the mean, median, third quartile, interquartile
range, and maximum value. This means that in the last movement, the difference between the
minimum and maximum acceleration values increased, translating into more impulsive
movement.

Lifting Stage - Elbows 3D

Acceleration
3D Acceleration (m/s2)

2,5

1,5

0,5

0
R Elbow
Figure 12 – Elbows LiftingI R
3DElbow F L Elbow I L Elbow
Acceleration
F

C. Shoulders

1) Shoulders 3D Position

As shown in Figure 13, the p-value (bilateral) for both shoulders were greater than the
significance level alpha = 0.05 (R Shoulder = 0.27 and L Shoulder = 0.11), meaning no
significant differences were observed in the shoulders position mean values.
The results showed an increase in 3D position values and it was mostly contributed by an
increase in the X and Y axis values, which correspond to a stooped position, both at the
beginning/end of the movement and at the maximum arm extension in the sagittal plane.

Figure 13 – Shoulders Student's paired t-test results

4. CONCLUSIONS
The questionnaires results showed that there is a slight tendency for people that practice
sports to have more endurance. On the other side, the lack of exercise and smoking might
harm the physical condition.
In order to compensate for muscle fatigue, people adapted their working strategy and
exhibited several muscle fatigue indicators. Males could handle more time performing the
experiment.
The statistical analysis showed that individuals increased their range on the X axis (increased
forward bending), both wrists increased their velocity in the lifting and lowering stage. In
terms of acceleration, half the participants had less fluid movement during lifting, with higher
acceleration values in some parts and lower values in others. Relatively to the elbows, subjects
closed the elbows and hold them near the body, in order to increase stability and hold the load
better. Volunteers performed the lifting with higher velocity in the final movement of the
experiment, especially on the left elbow. Both left wrist and elbow had higher differences from
beginning to end than the right side.

4.1. Strengths of the methodology

This methodology is a new approach that allows the assessment of common symptoms and
behaviors that people demonstrate when starting to feel muscle fatigue, which could be listed
and used as indicators in the work environment.
Motion analysis can be combined with accelerometry in order to assess muscle fatigue, by
noticing changes in accelerations and movement patterns, as well as angles between wrist,
shoulder, and elbow.
A specific checklist could be created for each type of work, by performing a pilot test with some
volunteers and analyzing the recordings. That way, a list of indicators could be used in the
future, to enable a quick assessment of the worker's physical condition.

4.2. Limitations of the methodology

The main limitation of this study is the fact that only the first and last movements of lifting and
lowering were studied, because the video analysis was made manually by marking each
reference point placed on the volunteer body in each frame of the video (1 frame = 0.03 s),
which made it impossible to assess the entire recordings. The gradual development of muscle
fatigue and registered symptoms were only described and not quantified.
The ideal solution would be the continuous analysis of the recordings, with automatic live
results, that allowed a fast assessment of the situation.
The analyzed sample of 13 people is not the ideal sample size to generalize the results and future
work should include a larger and representative sample.

REFERENCES
Al-Mulla, M.R, Sepulveda, F. & Colley M. (2011) A review of non-invasive techniques to detect and predict
localized muscle fatigue. Sensors, 11, pp. 3545-3594. http://dx.doi.org/10.3390/s110403545
Bogdanis, G. (2012) Effects of physical activity and inactivity on muscle fatigue. Frontiers in physiology.
Volume 3. Article 142. https://doi.org/10.3389/fphys.2012.00142
Brown, N., Bichler, S., Fiedler, M. & Alt, W. (2016). Fatigue detection in strength training using three-
dimensional accelerometry and principal component analysis. Sports Biomechanics , 15:2, pp. 139-
150. http://dx.doi.org/10.1080/14763141.2016.1159321
Enoka, R.M. & Duchateau, J., (2008). Muscle fatigue: what, why and how it influences muscle function. J.
Physiol. 586 (1), 11-23. http://dx.doi.org/10.1113/jphysiol.2007.139477
Gates, D. & Dingwell, J. (2011). The effects of muscle fatigue and movement height on movement
stability and variability. Experimental Brain Research , 209:4, pp. 525-536.
http://dx.doi.org/10.1007/s00221- 011-2580-8
Halim, I., Rawaida, R, K., A, R., Saptari, A. & Shahrizan, M. (2014). Analysis of muscle activity using
surface electromyography for muscle performance in manual lifting task. Applied Mechanics and
Materials, 564, pp. 644-649. http://dx.doi.org/10.4028/www.scientific.net/AMM.564.644
Hallowell, M. R. (2010). Worker fatigue: Managing concerns in rapid renewal highway construction projects.
Professional Safety, 55 (12), pp. 18-26.
Huang, X. & Hinze, J. (2006). Owner’s role in construction safety. J. Constr. Eng. Manage., pp. 164-173.
http://dx.doi.org/10.1061/(ASCE)0733-9364(2006)132:2(164)
George, D. & Mallery, M. (2010). SPSS for Windows Step by Step: A Simple Guide and Reference, 17.0
update (10a ed.) Boston: Pearson.
Lee, J., Nussbaum, M. & Kyung, G. (2014). Effects of work experience on fatigue-induced biomechanical
changes during repetitive asymmetric lifts/lowers. Ergonomics , 57:12, pp. 1875-
1885.http://dx.doi.org/10.1080/00140139.2014.957733
McDonald, A., Tse, C. & Keir, P. (2016). Adaptations to isolated shoulder fatigue during simulated
repetitive work. Part II: Recovery. Journal of Electromyography and Kinesiology , 29, pp. 42-
49.http://dx.doi.org/10.1016/j.jelekin.2015.05.005
Mehta, J., Lavender, S. & Jagacinski, R. (2014). Physiological and biomechanical responses to a prolonged
repetitive asymmetric lifting activity. Ergonomics , 57:4, pp. 575-588.
http://dx.doi.org/10.1080/00140139.2014.887788
Mehta, J., Lavender, S., Jagacinski, R. & Sommerich, C. (2015). Effects of Task Precision Demands on
Behavioral and Physiological Changes During a Repetitive Asymmetric Lifting Activity. Human Factors
and Ergonomics Society , 57:3, pp. 435-446. http://dx.doi.org/10.1177/0018720814551556
Oyewole, S. (2014). Enhancing Ergonomic Safety Effectiveness of Repetitive Job Activities: Prediction of
Muscle Fatigue in Dominant and Nondominant Arms of Industrial Workers. Human Factors and
Ergonomics in Manufacturing & Service Industries , 24:6, pp. 585-600.
http://dx.doi.org/10.1002/hfm.20590
Poole, R., Kennedy, O., Roderick, P., Fallowfield, J., Hayes, P., Parkes, J. (2017). Coffee consumption and
health: umbrella review of meta-analyses of multiple health outcomes British Medical Journal,.
https://doi.org/10.1136/bmj.j5024
Sluiter, J. K., De Croon, E. M., Meijman, T. F. & Frings-Dresen, M. H. W. (2003). Need for recovery from
work related fatigue and its role in the development and prediction of subjective health complaints.
Occup. Environ. Med., 60 (Suppl. 1), pp. 62-70. http://dx.doi.org/ 10.1136/oem.60.suppl_1.i62
Srinivasan, D. & Mathiassen, S.E. (2012). Motor variability in occupational health and performance.
Clinical Biomechanics. 27, pp. 979-993. http://dx.doi.org/10.1016/j.clinbiomech.2012.08.007
Toole, T. M. (2005). A project management causal loop diagram. Proc., 21st Conf. of Association of
Researchers in Construction Management (ARCOM 2005), Association of Researchers in Construction
Management (ARCOM), U.K., pp. 763-772.
Van Cutsem, J., Marcora, S., Pauw, K., Bailey, S., Meeusen, R. & Roelands, B. (2017). The effects of
mental fatigue on physical performance: A systematic Review. http://dx.doi.org/10.1007/s40279-016-
0672-0
Wüst, R., Morse, C., Haan, A., Rittweger, J., Jones, D., Degends, H. (2008) Skeletal muscle properties
and fatigue resistance in relation to smoking history. European Journal of Applied Physiology, pp. 103-
110. http://dx.doi.org/10.1007/s00421-008-0792-9