A CTA TECHNICA CORVINIENSIS – Bulletin of Engineering

Tome XI [2018] | Fascicule 2 [April – June]

1.
Ahmet SHALA, 2.Ramë LIKAJ

DYNAMIC MODELING OF 3 DoF ROBOT MANIPULATOR

1-2
University of Prishtina / Faculty of Mechanical Engineering, Prishtina, KOSOVO

Abstract: Dynamical Modeling of robots is commonly first important step of Modeling, Analysis and Control of robotic systems.
This paper is focused on using Denavit-Hartenberg (DH) convention for kinematics and Newton-Euler Formulations for dynamic
modeling of 3 DoF - Degree of Freedom of 3D robot. The process of deriving of dynamical model is done using Software Maple.
Derived Dynamical Model of 3 Dof robot is converted for Matlab software use for future analysis, control and simulations.
Keywords: Modelling, dynamics, robot, analysis

INTRODUCTION
Dynamics is a huge field of study devoted to studying the
forces required to cause motion. The dynamic motion of the
manipulator arm in a robotic system is produced by the
torques generated by the actuators. This relationship
between the input torques and the time rates of change of
the robot arm components configurations, represent the
dynamic modeling of the robotic system which is concerned
with the derivation of the equations of motion of the
manipulator as a function of the forces and moments acting
on. So, the dynamic modeling of a robot manipulator consists
of finding the mapping between the forces exerted on the
structures and the joint positions, velocities and accelerations.
A good model has to satisfy two conflicting objectives.
A robot manipulator is basically a positioning device. To
control the position we must know the dynamic properties of
the manipulator in order to know how much force to exert on
it to cause it to move: too little force and the manipulator is
slow to react; too much force and the arm may crash into
objects or oscillate about its desired position. Figure 1. Symbolic representation - Axes rotations for
Deriving the dynamic equations of motion for robots is not a Denavit-Hartenberg parameters
simple task due to the large number of degrees of freedom Denavit-Hartenberg transformation matrix for adjacent
and nonlinearities present in the system. This part is coordinate frames, i and i – 1.
concerned with the development of the dynamic model for cos(θ i ) − cos(α i ) ⋅ sin( θ i ) sin( α i ) ⋅ sin( θ i ) a i ⋅ ⋅cos(θ i )
3 Dof robot and their kinematics and dynamics equations.  sin( θ ) cos(α ) ⋅ cos(θ ) − sin( α ) ⋅ cos(θ ) a ⋅ ⋅sin( θ )
ROBOT STRUCTURE OF 3 DOF AND COORDINATE Ai =  i i i i i i i 

 0 sin( α i ) cos(α i ) di 
SYSTEMS  
Based on structure of 3 Dof robot (Figure 1), is created Table  0 0 0 1 
1 of Denavit-Hartenberg parameters for 3 DoF robot. Orthogonal rotation matrix Ri which transforms a vector in the
i-th coordinate frame to a coordinate frame which is parallel
Table 1. Denavit-Hartenberg parameters for 3 DoF robot to the (i-1)-th coordinate frame is first 3x3 sub-matrices of Ai:
Link # ai αi di θi
*
cos(θ i ) − cos(α i ) ⋅ sin( θ i ) sin( α i ) ⋅ sin( θ i ) 
q1 R i =  sin( θ i ) cos(α i ) ⋅ cos(θ i ) − sin( α i ) ⋅ cos(θ i )
1 0 0 d1
2 0 −β d2
*
0  0 sin( α i ) cos(α i ) 
*
3 0 0 L 3a q3 for i=1,2, … , N, where R N +1 = E = diag(1)
* Joint variable DYNAMIC EQUATIONS – NEWTON-EULER FORMULATION
Dynamics of robot is the study of motion with regard to forces
(the study of the relationship between forces/torques and

95 | F a s c i c u l e 2
A CTA TECHNICA CORVINIENSIS – Bulletin of Engineering
Tome XI [2018] | Fascicule 2 [April – June]

motion). A dynamic analysis of a manipulator is useful for the  T

following purposes: R ⋅ v + ω  × p + ω × (ω × p )
 i +1 i
— It determines the joint forces and torques required to 
i +1 i +1 i +1 i +1 i +1

produce specified end-effector motions (the direct if joint is rotational

dynamic problem). 

— It produces a mathematical model which simulates the 
motion of the manipulator under various loading 
  
conditions (the inverse dynamic problem) and/or control  T  (3)
v = R v + z d + 2ω x(z d ) +
 
schemes.  
  
— It provides a dynamic model for use in the control of the i +1
 i +1  i 0 i +1 i 0 i +1 

actual manipulator. 
+ ω xp + ω x(ω xp )
Dynamic modelling of mechanical structures can be a  i +1 i +1 i +1 i +1 i +1
complex problem. In robotics, more specifically, in 
manipulators, there are two methodologies used for dynamic 
if joint is translational
modelling. 
NEWTON-EULER FORMULATION 

The Newton-Euler formulation [1] shown in equations (1)-(9) where: p i = [a i d i ⋅ sin( α i ) d i ⋅ sin( α i )] is position of
T

computes the inverse dynamics (ie., joint torques/forces from the i-th coordinate frame with respect to the (i – 1)-th
joint positions, velocities, and accelerations) bases on two coordinate frame.
sets of recursions: the forward and backward recursions. The
Initial conditions: ω0 = ω 0 = v 0 = 0 ; Gravitational
forward recursions (1)-(3) transform the kinematics variables
from the base to the end-effector. The initial conditions (for acceleration: v 0 = g x [
gy gz .
T
]
i=0) assume that the manipulator is at rest in the gravitational Linear acceleration of the center-of-mass of link i
field. The backward recursions (4)-(9) transform the forces and ai = ω
 i × s i + ωi × (ωi × s i ) + v i (4)
moments from the end-effector to the base, and culminate
where: si is position of center-of-mass of link i
with the calculation of the joint torques/forces.
Net force exerted on link i:
Angular velocity of the i-th coordinate frame
Fi = m i ⋅ a i (5)
  
R T ⋅ ω + z ⋅ θ  if joint is rotational Net moment exerted on link i
   N i = Ii ω
 i + ω i × (I i ⋅ ω i ) (6)
 i +1  i 0 
i +1 
I ixx 0 0
ω = (1)
 T
where I =  0 I iyy 0 
i +1 R ⋅ ω if joint is translational i 

 i +1 i  0 0 I izz 
 Ii is moment of inertia tensor of link i about the centre-of-
where: z 0 = [0 0 1]
T mass of link i (parallel to the i-th coordinate frame), with only
principal inertias Iixx, Iiyy and Iizz. Because of symmetry of link
Angular acceleration of the i-th coordinate frame frames, cross-inertias can be used zero.
   Force exerted on link i by link i – 1:
R T ⋅ ω + z ⋅ θ + ω × (z ⋅ θ )
  f i = R iT+1 ⋅ f i+1 + Fi (7)
  
 i+1  i 0 i +1 i 0 i +1 
 Moment exerted on link i by link i – 1
 n i = R Ti+1 ⋅ n i+1 + p i × f i + N i + s i × Fi (8)
 if joint is rotational
 Joint torque/force at joint i:
 n ⋅ (R T ⋅ z ) if joint is rotational
ω
 = (2) 
R T ⋅ ω
  iT

i +1 0
i +1
 τ = (9)
 i +1 i  f ⋅ (R T ⋅ z ) if joint is translational
 i
 iT i +1
 if joint is translational 
0


 DERIVING OF DYNAMICAL MODEL FOR 3 DOF ROBOT
 Based on Newton-Euler formulation (1-9), rotation matrices
Linear acceleration of the i-th coordinate frame for links of robot (i=1,2,3) are:

96 | F a s c i c u l e 2
 A CTA TECHNICA CORVINIENSIS – Bulletin of Engineering
Tome XI [2018] | Fascicule 2 [April – June]

cos(q 1 ) − sin(q 1 ) 0 I3xx 0 0 

R 1 =  sin(q 1 ) cos(q 1 ) 0 I3 =  0 I3yy 0 
 0 0 1  0 0 I3zz 
1 0 0  Net force and moment exerted on link i = 1,2,3:

R 2 = 0 cos(β) sin(β) F1 = m 1 ⋅ a 1
0 − sin(β) cos(β) F2 = m 2 ⋅ a 2
cos(q 3 ) − sin(q 3 ) 0 F3 = m 3 ⋅ a 3
R 3 =  sin(q 3 ) cos(q 3 ) 0 N 1 = I1 ⋅ ω  1 × I1 ⋅ ω
 1 + (ω  1)
 0 0 1 N 2 = I2 ⋅ ω
 2 + (ω
 2 × I2 ⋅ ω
 2)
Initial conditions are: N 3 = I3 ⋅ ω
 3 + (ω
 3 × I3 ⋅ ω
 3)
0 0 0 0 0 Backward recursion (i=3,2,1); Force and moment exerted on
     0 = 0 , v 0 = 0 , v 0 = 0
z 0 = 0 , ω0 = 0 , ω    
      link i by link i-1.
1 0 0 0 0 By supposing that there are no outside load for end-effector
are:
Forward recursions, for 3 DoF robot, (i=1,2,3).
Angular velocity of the i-th coordinate frame:
0 0 1 0 0
T
ω1 = R 1 ⋅ (ω0 + z 0 ⋅ q 1 )
 
f4 = 0 , n 4 = 0 , R 4 = 0 1 0
 

T 0 0 0 0 1

ω 2 = R 2 ⋅ ω1
T For link #3:
ω3 = R 3 ⋅ (ω2 + z 0 ⋅ q 3 ) f3 = R 4 ⋅ f4 + F3
Angular acceleration of the i-th coordinate frame:
n 3 = R 4 ⋅ n 4 + p3 × f3 + N 3 + s 3 × F3
 1 = R 1 T ⋅ (ω
ω  0 + z0 ⋅ q
 1 ) + ω0 × z 0 ⋅ q 1
Torque at joint 3:
T
ω
 2 = R2 ⋅ ω
1 T
τ 3 = n 3 ⋅ (R 3 ⋅ z 0 )
 3 = R 3 T ⋅ (ω
ω  2 + z0 ⋅ q
 3 ) + ω2 × z 0 ⋅ q 3 For link #2:
Position of the i-th coordinate frame with respect to the (i- f2 = R 3 ⋅ f3 + F2
1)-th coordinate frame: n 2 = R 3 ⋅ n 3 + p2 × f2 + N 2 + s 2 × F2
0 0  0  Force at joint 2:
p1 =  0  , p 2 =  0  , p3 =  0 
    T
τ 2 = f 2 ⋅ (R 2 ⋅ z 0 )
L 1  d 2  L 2b  For link #1:
Linear acceleration of the i-th coordinate frame: f1 = R 2 ⋅ f2 + F1
T
v 1 = R 1 ⋅ v 0 + ω
 1 × p1 + ω
 1 × (ω
 1 × p1 ) n 1 = R 2 ⋅ n 2 + p1 × f1 + N 1 + s 1 × F1
T
 1 × z 0 ⋅ q 2 ) + (ω1 × p1 )
v 2 = R 2 ⋅ v 1 + z 0 ⋅ q 2 + (2ω Torque at joint 1:
T
T
 3 × p3 + ω
v 3 = R 3 ⋅ v 2 + ω  3 × (ω
 3 × p3 ) τ 1 = n 1 ⋅ (R 1 ⋅ z 0 )
Linear acceleration of the center-of-mass of link i: In the end the vector of Forces-Torques for 3 DoF robot is:
a 1 = ω 1 × s 1 + ω1 × (ω1 × s 1 ) + v 1 τ1 
 2 × s 2 + ω2 × (ω2 × s 2 ) + v 2
a 2 = ω τ = τ 2 
a 3 = ω 3 × s 3 + ω3 × (ω3 × s 3 ) + v 3 τ3 
Moment of inertia tensor of link i about the center of mass of Modeling-Calculations of 3 DoF robot is done using Maple
link i (parallel to the i-th coordinate frame), with only principal software, equations are converted for Matlab use.
inertias Iixx, Iiyy and Iizz. Because of symmetry of link frames, CONCLUSION
cross-inertias are used zero. Based on presented paper can be concluded that Newton-
I1xx 0 0  I2xx 0 0  Euler formulation is very useful for dynamical modeling of
  systems generally and robotic systems especially.
I1 =  0 I1yy 0  , I2 = 0 I2yy 0  ,

  Use of Maple software is very useful for modelling of complex
 0 0 I1zz   0 0 I 2zz  systems and representations of results symbolically –

97 | F a s c i c u l e 2
A CTA TECHNICA CORVINIENSIS – Bulletin of Engineering
Tome XI [2018] | Fascicule 2 [April – June]

representation of expressions of dynamical model with many

characters (until 100000).
Opportunity of Maple software to convert derived
expressions for Matlab use is very helpful for future analyses
and simulations of systems.
Acknowledgements
The first author is profoundly thankful to the corresponding
author Rame Likaj (email: [email protected]) which has paid
attention to fulfill all requirements about this research work.
References
[1] Luh, J. Y. S., Walker, M. W. and Paul, R. P.; On-Line
Computational Scheme for Mechanical Manipulators.
Journal of Dynamic Systems, Measurement, and Control
102(2):69-76, June, 1980.
[2] Brady, M., et al. (editors). Robot Motion: Planning and
Control. MIT Press, Cambridge, MA, 1982.
[3] Khosla, Pradeep, "Estimation of Robot Dynamics
Parameters: Theory and Application". Institute for Software
Research. Paper 651. http://repository.cmu.edu/isr/651,
Carnegie Mellon University, 1987.
[4] A. Shala, R. Likaj; Design of Genetic Algorithm for
optimization of Fuzzy Neural Network Controller, 8th
International Conference Modern Technologies in
Manufacturing, Cluj-Napoca, Romania, 2007.
[5] H.J. Sommer: Vector loops facilitate simple kinematic
analysis of planar mechanisms, either closed-chain or open-
chain, Mechanical & Nuclear Engineering Faculty, USA,
www.mne.psu.edu/sommer/me50/
[6] J. M. McCarthy and G. S. Soh, 2010, Geometric Design of
Linkages, Springer, New York.
[7] B. Paul, Kinematics and Dynamics of Planar Machinery,
Prentice-Hall, NJ, 1979
[8] L. W. Tsai, Robot Analysis: The mechanics of serial and
parallel manipulators, John-Wiley, NY, 1999.
[9] K. J. Waldron and G. L. Kinzel, Kinematics and Dynamics, and
Design of Machinery, 2nd Ed., John Wiley and Sons, 2004.
[10] Torby, Bruce (1984). "Energy Methods". Advanced
Dynamics for Engineers. HRW Series in Mechanical
Engineering. United States of America: CBS College
Publishing. ISBN 0-03-063366-4.

ISSN: 2067-3809
copyright © University POLITEHNICA Timisoara,
Faculty of Engineering Hunedoara,
5, Revolutiei, 331128, Hunedoara, ROMANIA
http://acta.fih.upt.ro

98 | F a s c i c u l e 2