1212-ArticleText-1921-1-10-20191031

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/336956363
Dynamics and Computed-Torque Control of a 2-DOF manipulator:

Mathematical Analysis
Article in International Journal of Advanced Science and Technology · October 2019
CITATIONS READS
24 2,679
2 authors:
Abdel-Nasser Sharkawy Panagiotis Koustoumpardis

South Valley University University of Patras
67 PUBLICATIONS 790 CITATIONS 48 PUBLICATIONS 623 CITATIONS
SEE PROFILE SEE PROFILE
All content following this page was uploaded by Abdel-Nasser Sharkawy on 01 November 2019.
The user has requested enhancement of the downloaded file.

International Journal of Advanced Science and Technology
Vol. 28, No. 12, (2019), pp. 201-212
Dynamics and Computed-Torque Control of a 2-DOF manipulator:

Mathematical Analysis
Abdel-Nasser Sharkawy1, 2*, Panagiotis N. Koustoumpardis2

1
Mechanical Engineering Department, Faculty of Engineering, South Valley University,
Qena 83523, Egypt
2
Department of Mechanical Engineering and Aeronautics, University of Patras, Rio 26504,
Greece
Abstract
In this paper, the dynamics of the 2-DOF manipulators are presented. The mathematical analysis
for these dynamics for two cases is carried out: in the first case, the manipulator is moving without
any collision (external forces) with its environment, whereas the second case the manipulator is
collided. The Computed-Torque Controlis usedfor these dynamic of the manipulator. A simulation
study is executed using a sinusoidal motion commanded simultaneously to the two joints of the
manipulator. The actual and desired signals of the joints’positions, velocities, accelerations and
torques of the 2-DOF robotare compared whether there is a collision or no.The results prove that the
computed-torque control is effectivelyminimizing the error between the actual signals and the desired
signals. In addition, the dynamic coupling between the joints is presented from the results.
Keywords:Robot-Manipulator, Joints Dynamics, Dynamic Coupling, Collision, Computed-Torque

Control.
1. Introduction
The robot manipulators are highly nonlinear, dynamically coupled and time-varying systems which
are used extensively in industrial applications [1].The benefit of the dynamic model of the
manipulator is to compute the torque and force required in order to execute the typical work cycle and
to give vital information for the design of the links, the joints, the drives, and the actuators as well as
for the control scheme. The manipulator dynamic behaviour gives a relationship between the joint
actuator torques and the links motion for simulation and design of control algorithms. Dynamics
analysis of robot manipulatoris investigated by many researchers in[2]–[5] and the dynamic coupling
between the joints of the manipulator is investigated experimentally in [6].
Computed-Torque Control is an effective motion control strategy for robotic manipulator
systems[7]. Computed Torque Control [8] is worth noting because of the easiness to be understood
and of its good performances.Computed Torque Control is a method for linearizing and decoupling
the robotic dynamics by using perfect dynamical models of robotic manipulator systems in order to
each joint motion can individually be controlled using other well-developed linear control
strategies[7].Computed-Torque controller was proposed by many researchersfor a parallel
manipulator[9],for a master-slave robot manipulator system[10], andstable computed-torque control
was proposed in [11] of robot manipulators via fuzzy self-tuning.
ISSN: 2005-4238 IJAST 201

Copyright ⓒ 2019 SERSC
Vol. 28, No. 12, (2019), pp. 201-212
In this paper, the mathematical analysisfor the dynamics of 2-DOF Manipulator is presented.The
computed-torque control law is used for these dynamics.The dynamic coupling between the
manipulator joints is investigated and presented.Indeed, the main benefit from this procedure is that
the dynamic coupling between the joints should be taken into account during the design and the
implementation of the human-robot collision detection method and the collided link identification
which contribute to the safety of the human-robot interaction.
2. Dynamics of Multiple Joints Motion

For the presented case study, it is assumed that the commanded motion is applied to the 4th (A3)
and 6th (A5) joint of the KUKA LWR manipulator (collaborative robot), as shown in Fig. 1. The case
when the two joints during the planar horizontal motion is studied and analyzed (Fig.1b). In all the
paper, joint 1 and joint 2 will representA3 andA5 joints respectively. It should be noted that the all
other joints of the manipulator are fixed.
+ 𝑟2
𝜃2
𝐿1
Joint 2 (A5)
𝑟1
𝑦
𝜃1
𝑥
Joint 1 (A3)
(a) (b)
Figure 1. Kuka LWR manipulator. (a) The all joints of the kuka robot.(b) Motion of Joint A3 and A5
(The black spot means the center of mass) in a horizontal plane.
The second-order vector differential equation for the motion of the manipulator as a function of the
applied joint torques when there is no external forces (collisions) affecting on the end-effector or the
link between the two joints is given by [12]
𝑀 𝜃 𝜃 + 𝐶 𝜃, 𝜃 𝜃 = 𝜏 (1)
In another form, the equation can be rewritten as[12]
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝜏1

+ = 𝜏 (2)
𝛿 + 𝛽𝑐2 𝛿 𝜃2 𝛽𝑠2 𝜃1 0 𝜃2 2
where
𝛼 = 𝐼𝑧1 + 𝐼𝑧2 + 𝑚1 𝑟12 + 𝑚2 𝐿21 + 𝑟22
𝛽 = 𝑚2 𝐿1 𝑟2
ISSN: 2005-4238 IJAST 202

Vol. 28, No. 12, (2019), pp. 201-212
𝛿 = 𝐼𝑧2 + 𝑚2 𝑟22
where𝜃𝑖 , 𝜃𝑖 , 𝜃𝑖 is the actual angular position of the center of mass for the ith link and their
corresponding time derivatives.𝑟𝑖 is the distance from the joint 𝑖 to the center of mass for link𝑖.𝐼𝑧𝑖 is
the moment of inertia about the z-axes of the ith link frame.𝑀 𝜃 ∈ 𝑅 2×2 is the inertia matrix that
depends on thevariable𝜃, and𝐶 𝜃, 𝜃 ∈ 𝑅 2×2 is the matrix containing the Coriolis and centrifugal
terms.𝐺 𝜃 ∈ 𝑅 2 = 0.0is the gravity vector and it is equal to zero due to that the motion in horizontal
plane. 𝜏𝑖 ∈ 𝑅 2 is the joint torque for ith joint.𝑚𝑖 is the mass of ith link, 𝐿𝑖 is the length of ith link,
𝑠𝑖 = 𝑠𝑖𝑛𝜃𝑖 , 𝑐𝑖 = 𝑐𝑜𝑠𝜃𝑖 , 𝑠𝑖𝑗 = 𝑠𝑖𝑛(𝜃𝑖 + 𝜃𝑗 ), and 𝑐𝑖𝑗 = 𝑐𝑜𝑠(𝜃𝑖 + 𝜃𝑗 ).
From (2), it is easy to derive the torque at joint1 and 2 as the following equations
𝜏1 = 𝛼 + 2𝛽𝑐2 𝜃1 + 𝛿 + 𝛽𝑐2 𝜃2 − 𝛽𝑠2 𝜃1 𝜃2 − 𝛽𝑠2 (𝜃1 + 𝜃2 )𝜃2 (3)
𝜏2 = 𝛿 + 𝛽𝑐2 𝜃1 + 𝛿𝜃2 + 𝛽𝑠2 𝜃12 (4)
2.1. External Force is perpendicular to link 1 (𝐋𝟏 )
If there is an external force exerted perpendicularly on the link 1 as shown in Fig. 2, the dynamic
equation is derived by the following steps,
𝐹 + 𝑟2
𝑟1 𝜃2
𝑑𝜃1
𝜃1 𝐿1
Joint 2
𝐹
𝑟1
𝑦
𝜃1
𝑥 𝐹
Joint 1
(a) (b)
Figure 2. Effect of external force (collision) on the joints torque.(a) Link after rotating by the effect of
the external force. (b) External perpendicular force (collision) on link 1.
The work done by force 𝐹 is given by
𝑑𝑊 = 𝐹𝑟1 𝑑𝜃1 (5)
ISSN: 2005-4238 IJAST 203

Vol. 28, No. 12, (2019), pp. 201-212
The potential energy 𝑉 𝜃 is equal to the work done by force 𝐹 so
𝑉 𝜃 = 𝑑𝑊 = 𝐹𝑟1 𝑑𝜃1 (6)

Taking the derivative of (6) according to𝜃1 and 𝜃2
𝜕𝑉 𝜕𝑉
𝜕𝜃1
= 𝐹𝑟1 , 𝜕𝜃2
= 0(7)
So the equation for the motion of the manipulator is converted from (2) into the following equation
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝐹𝑟1 𝜏1

+ + = 𝜏 (8)
𝛿 + 𝛽𝑐2 𝛿 𝜃2 𝛽𝑠2 𝜃1 0 𝜃2 0 2
It is noted from (8) that the torque at joint 1 is more affected than the torque at joint 2.
2.2. External Force is perpendicular to link 2 (𝐋𝟐 )
In case of the perpendicular external force on link 2 as shown in Fig. 3, the dynamic equation is
derived as the following steps,
𝑟2 𝜃2
𝐹
𝑎 Joint 2
𝑦
𝜃1
𝑥
Joint 1
𝑟1
𝐿1
Figure 3. External perpendicular force (collision) exerted on link 2
The potential energy 𝑉 𝜃 is equal to the work done by force 𝐹
𝑉 𝜃 = 𝑑𝑊1 + 𝑑𝑊2 (9)
The work done by force 𝐹 at joint 2
𝑑𝑊1 = 𝐹𝑟2 𝑑(𝜃1 + 𝜃2 )(10)
ISSN: 2005-4238 IJAST 204

Vol. 28, No. 12, (2019), pp. 201-212
where
𝑑 𝜃1 + 𝜃2 = 𝜃1 + 𝜃2 𝑎𝑓𝑡𝑒𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑛𝑔 𝑡𝑕𝑒 𝑙𝑖𝑛𝑘 − 𝜃1 + 𝜃2 𝑏𝑒𝑓𝑜𝑟𝑒 𝑟𝑜𝑡𝑎𝑡𝑖𝑛𝑔 𝑡𝑕𝑒 𝑙𝑖𝑛𝑘
The work done by force F at joint 1
𝑑𝑊2 = 𝐹 𝑟2 + 𝑎 𝑑 𝜃1 + 𝜃2 = 𝐹 𝑟2 + 𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2
= 𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2 (11)
By substituting (10) and (11) into (9), the potential energy is
𝑉 𝜃 = 𝐹𝑟2 𝑑(𝜃1 + 𝜃2 ) + 𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2
= 2𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2 (12)
Taking the derivative of (12) according to𝜃1 and 𝜃2
𝜕𝑉 𝜕𝑉
𝜕𝜃1
= 2𝐹𝑟2 + 𝐹𝐿1 cos 𝜃2 , 𝜕𝜃2
= 2𝐹𝑟2 + 𝐹𝐿1 cos 𝜃2 − 𝐹𝐿1 sin 𝜃2 𝑑 𝜃1 + 𝜃2 13)
So the dynamic equation for the motion of the manipulator is converted from (2) into the following
equation
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1

+
𝛿 + 𝛽𝑐2 𝛿 𝜃2 𝛽𝑠2 𝜃1 0 𝜃2
2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2 𝜏1
+ = 𝜏
2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2 − 𝐹𝐿1 𝑠𝑖𝑛 𝜃2 𝑑 𝜃1 + 𝜃2 2
(14)
It is noted from eq. (14) that the torque at joint 1 is more affected than joint 2 because of the
higher lever-arm.
Note: To derive the dynamic equation when there is an external force on link 1 and in the same time
another force on link 2, the same previous steps can be followed.
3. Computed Torque Control

Computed torque is a special application of feedback linearization of nonlinear systems, which has
gained popularity in modern systems theory [13], [14]. Computed-Torque Control is used in this
paper to simulate the reality as working with a real robot to show the effects on the variables;
position, velocity, acceleration and torque whether there is external force or no. The following steps
derive the equation of the computed-Torque Control law.
If the desired trajectory 𝜃𝑑 (𝑡) is selected for the arm manipulation so the tracking error is
𝑒 𝑡 = 𝜃𝑑 𝑡 − 𝜃(𝑡)(15)
ISSN: 2005-4238 IJAST 205

Vol. 28, No. 12, (2019), pp. 201-212
By taking the first and second derivative of the error is
𝑒 = 𝜃𝑑 − 𝜃 , 𝑒 = 𝜃𝑑 − 𝜃 (16)
From (1), 𝜃 is given by
𝜃 = 𝑀(𝜃)−1 𝜏 − 𝐶 𝜃, 𝜃 𝜃 (17)
By substituting 𝜃 from (17) into (16) so
𝑒 = 𝜃𝑑 + 𝑀(𝜃)−1 𝐶 𝜃, 𝜃 𝜃 − 𝜏 (18)
Defining the control input function as [15]
𝑢 = 𝑒 = 𝜃𝑑 + 𝑀(𝜃)−1 𝐶 𝜃, 𝜃 𝜃 − 𝜏 (19)
From (19), the computed joint torque 𝜏 is given by
𝜏 = 𝑀 𝜃 𝜃𝑑 − 𝑢 + 𝐶 𝜃, 𝜃 𝜃(20)
Select the control signal 𝑢 as the proportional-Derivative (PD) Feedback
𝑢 = −𝐾𝑑 𝑒 − 𝐾𝑝 𝑒(21)
By substituting (21) into (20), the computed joint torque which is the robot arm input becomes
𝜏 = 𝑀 𝜃 𝜃𝑑 + 𝐾𝑑 𝑒 + 𝐾𝑝 𝑒 + 𝐶 𝜃, 𝜃 𝜃(22)
Which called the computed-Torque Control law. The PD gains are usually selected for critical
damping𝜉 = 1[15], [16]. In this case:
𝐾𝑑2
𝐾𝑑 = 2 𝐾𝑝 and 𝐾𝑝 = 4
(23)
The computed-torque control depends on the inversion of the robot dynamics, and indeed is called
inverse dynamics control and it results the real joint acceleration vector after calculating 𝜏 from eq.
(22) and substituting its value in eq. (17).
The equation for the entire system can be derived from (1) and (22) by
𝑀 𝜃 𝜃 + 𝐶 𝜃, 𝜃 𝜃 = 𝑀 𝜃 𝜃𝑑 + 𝐾𝑑 𝑒 + 𝐾𝑝 𝑒 + 𝐶 𝜃, 𝜃 𝜃
𝜃 = 𝜃𝑑 + 𝐾𝑑 𝑒 + 𝐾𝑝 𝑒
Therefore,
𝑒 + 𝐾𝑑 𝑒 + 𝐾𝑝 𝑒 = 0.0(24)
Equation (24) means that there is not any external disturbance. The block diagram of the computed–
torque control is shown in Fig. 4
ISSN: 2005-4238 IJAST 206

Vol. 28, No. 12, (2019), pp. 201-212
Figure 4. Computed-torque control scheme. In our case,𝐺 𝜃 = 0.0 because of the motion is in
horizontal plane.
3.1. Computed Torque Control for Multiple joints motion
The torque from the PD-controller based on Fig. 4 is given by the following equation
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1𝑑 + 𝐾𝑑1 𝑒1 + 𝐾𝑝1 𝑒1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝜏𝑐1

+ = 𝜏 (25)
𝛿 + 𝛽𝑐2 𝛿 𝜃2𝑑 + 𝐾𝑑2 𝑒2 + 𝐾𝑝2 𝑒2 𝛽𝑠2 𝜃1 0 𝜃2 𝑐2
A. If there is no external force

When there are no external disturbances, the computed joint torque is equal to the torque resulted
from the controller as
𝜏1 = 𝜏𝑐1 and 𝜏2 = 𝜏𝑐2 (26)
By substituting (26) into (25) and make the equality with (2), the equation for the entire system is
derived as
𝑒1 + 𝐾𝑑1 𝑒1 + 𝐾𝑝1 𝑒1
= 0.0(27)
𝑒2 + 𝐾𝑑2 𝑒2 + 𝐾𝑝2 𝑒2
B. If there is external force that is perpendicular to link 1

According to the equations (5-8) and (25), the computed joint torques for the two joints are derived
as
𝜏1 𝜏𝑐1 𝐹𝑟1
𝜏2 = 𝜏𝑐2 + 0 (28)
By substituting (25) into (28) and make the equality with (2), then the equation for the entire system is
derived as
ISSN: 2005-4238 IJAST 207

Vol. 28, No. 12, (2019), pp. 201-212
𝑒1 + 𝐾𝑑1 𝑒1 + 𝐾𝑝1 𝑒1 −1
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝐹𝑟1
=− (29)
𝑒2 + 𝐾𝑑2 𝑒2 + 𝐾𝑝2 𝑒2 𝛿 + 𝛽𝑐2 𝛿 0
C. If there is external force that is perpendicular to link 2

From equations (9-14) and (25), the computed joint torques for the two joints are calculated by
𝜏1 𝜏𝑐1 2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2
𝜏2 = 𝜏𝑐2 + 2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2 − 𝐹𝐿1 𝑠𝑖𝑛 𝜃2 𝑑 𝜃1 + 𝜃2 (30)
By substituting (25) into (30) and make the equality with (2), then the equation for the entire system is
derived as
𝑒1 + 𝐾𝑑1 𝑒1 + 𝐾𝑝1 𝑒1
𝑒2 + 𝐾𝑑2 𝑒2 + 𝐾𝑝2 𝑒2
−1
𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2
= −
𝛿 + 𝛽𝑐2 𝛿 2𝐹𝑟2 + 𝐹𝐿1 𝑐𝑜𝑠 𝜃2 − 𝐹𝐿1 𝑠𝑖𝑛 𝜃2 𝑑 𝜃1 + 𝜃2
(31)
After calculating the joint torques 𝜏1 and 𝜏2 whether there is disturbance or no, then substituting their
values into (2) to calculate the actual acceleration as
𝜃1 𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 −1 𝜏1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1
= 𝜏 − (32)
𝜃2 𝛿 + 𝛽𝑐 2 𝛿 2 𝛽𝑠2 𝜃1 0 𝜃2
By make the integral for 𝜃1 and𝜃2 , the actual velocity and position are determined.
4. Simulation Study
A program in Matlab is implemented based on the computed torque control for the multiple joints
motion to simulate the reality and show the differences between the actual and the reference positions,
velocities, accelerations, and torques of the two joints whether there is external force or no.
4.1. Determining the Parameters
The parameters used during the simulation are real data and collected from the KUKA LWR robot
datasheet [17]as shown in table I. A sinusoidal motion is commanded to the two joints as
𝜃𝑑 𝑡 = 𝐴 − 𝐴𝑐𝑜𝑠(2𝜋𝑓𝑡)(33)
𝜋
where 𝐴 is therange of the sinusoidal motion in radians= 4 ,
𝑓 is the frequency of the sinusoidal motion.
Using Zeigler-Nichols rules, the PD gains that give the best performance are given as
𝐾𝑝1 = 4,𝐾𝑑1 = 4, 𝐾𝑝2 = 4, 𝐾𝑑2 = 4
Table I. The values of the parameters used in the simulation[17]
ISSN: 2005-4238 IJAST 208

Vol. 28, No. 12, (2019), pp. 201-212
Parameter Value
𝐿1 0.39 𝑚
𝑟1 0.195 𝑚
𝑟2 0.078 𝑚
𝑚1 3.3 𝑘𝑔
𝑚2 0.3 𝑘𝑔
2
𝐼𝑧1 𝑚1 𝑟1 ≈ 0.1255 𝑘𝑔. 𝑚2
𝐼𝑧2 𝑚2 𝑟22 ≈ 0.00183 𝑘𝑔. 𝑚2
4.2. The Results
The diagrams of the actual and desired joints positions, velocities, accelerations and the computed
joint torques 𝜏1 and 𝜏2 whether there is external force or no are shown from Fig. 5 to Fig. 7.
(a) Joint 1 (b) Joint 2

Figure 5. The actual and desired joints’ positions, velocities, accelerations and the computed torques
when there are no external forces.
ISSN: 2005-4238 IJAST 209

Vol. 28, No. 12, (2019), pp. 201-212

Figure 6. The actual and desired joints’ positions, velocities, accelerations and the computed torques when there
is external force perpendicular to link 1 and equal to 10 N at time = 5 sec.

Figure 7. The actual and desired joints’ positions, velocities, accelerations and the computed torques when there
is external force perpendicular to link 2 and equal to 10 N at time = 5 sec.
ISSN: 2005-4238 IJAST 210

Vol. 28, No. 12, (2019), pp. 201-212
As shown from the resultsthe actual joint position, velocity and acceleration are approximately
coincide with the desired (reference) signals (Fig. 5) because of using the PD-controller. The errors
between theactual signals and desired ones are very small (neglected), therefore, we can conclude that
the PD-controller works very well and is a robust controller.
When there is external force or disturbance, the actual signal is affected by this force at the time
when the force exerts on the link (time = 5 s in Fig.6 and Fig. 7) then the PD-controller tries to
coincide again the actual signalwith the desired signal after the force effect finishes.
The dynamic coupling between the joints of the manipulators are presented from the figures (Fig. 6
and Fig. 7) and discussed as following:
 When there is external force (or in another words collision) whether on link 1 or link 2, these
variables; position, velocity and acceleration of joint 2 are more affected than of joint 1.
When the force exerts on link 2 the effect of this force on these variables is more than when it
exerts on link 1 and this is observed also from the mathematical calculations.
 The computed torque of joint 1 is higher than of joint 2 in any case whether there is collision
or no. In the case where the force is perpendicular to link 1, there is no change on the
computed torque of joint 2 (Fig. 6b) whereas the computed torque of joint 1 is affected at the
time when the force exerts on the link (Fig. 6a). When the collision on link 2, both the
computed torques are affected at the time when the force exerts on the link whereas the
computed torque of joint 1 is affected more than of joint 2 because of the higher lever-arm
(Fig. 7).
5. Conclusions
In this paper, themathematical modelling of the 2-DOF manipulator dynamics is presented. The
analysis is presented in two cases; the first onewithout collisions applied to the links of the
manipulator, whereas the second one is the collided applied case. The Computed-Torque Control is
also presented and a simulation study is executed using a sinusoidal motion commanded
simultaneously to the two joints of the manipulator. The results prove that the computed-torque
control is effectively minimizing the error between the actual signals and the desired signals of the
joints’ positions, velocities, accelerations and torques of the manipulator. Moreover, the dynamic
coupling between the joints is presented. The dynamic coupling between the manipulator joints
should be considered during the design and implementing the human-robot collision detection method
and collided link identification which contribute to the safety of the human-robot interaction.
References
[1] J. Shah, S. S. Rattan, and B. C. Nakra, “DYNAMIC ANALYSIS OF TWO LINK ROBOT
MANIPULATOR FOR CONTROL DESIGN USING COMPUTED TORQUE CONTROL,” Int. J. Res.
Comput. Appl. Robot., vol. 3, no. 1, pp. 52–59, 2015.
[2] R. Di Gregorio, “Dynamic model and performances of 2-DOF manipulators,” Robotica, vol. 24, no.
2006, pp. 51–60, 2019.
ISSN: 2005-4238 IJAST 211

Vol. 28, No. 12, (2019), pp. 201-212
[3] D. Naderi, A. Meghdari, and M. Durali, “Dynamic modeling and analysis of a two d . o . f . mobile
manipulator,” Robotica, vol. 19, pp. 177–185, 2001.
[4] H. Høifødt, “Dynamic Modeling and Simulation of Robot Manipulators,” Norwegian University of
Science and Technology Department, 2011.
[5] M. Shahab, 2DOF Robotic Manipulator: Control Design & Simulation, vol. 20, December. 2008.
[6] A. Sharkawy, P. N. Koustoumpardis, and N. Aspragathos, “Human – robot collisions detection for safe
human – robot interaction using one multi-input – output neural network,” Soft Comput., vol. 7, 2019.
[7] Z. Song, J. Yi, D. Zhao, and X. Li, “A computed torque controller for uncertain robotic manipulator
systems : Fuzzy approach,” Fuzzy Sets Syst., vol. 154, pp. 208–226, 2005.
[8] R. H. MIDDLETON and G. C. GOODW1N, “Adaptive computed torque control for rigid link
manipulations,” Syst. Control Lett., vol. 10, pp. 9–16, 1988.
[9] Z. Yang, J. Wu, J. Mei, J. Gao, and T. Huang, “Mechatronic Model Based Computed Torque Control of
a Parallel Manipulator,” Int. J. Adv. Robot. Syst., vol. 5, no. 1, pp. 123–128, 2008.
[10] O. O. Obadina, M. Thaha, K. Althoefer, and M. H. Shaheed, “A Modified Computed Torque Control
Approach for a Master-Slave Robot Manipulator System,” in Towards Autonomous Robotic Systems.
TAROS 2018. Lecture Notes in Computer Science, vol. vol 10965, M. Giuliani, T. Assaf, and M.
Giannaccini, Eds. Springer, Cham, 2018, pp. 28–39.
[11] M. A. Llama, R. Kelly, and V. Santibañez, “Stable Computed-Torque Control of Robot Manipulators
via Fuzzy Self-Tuning,” IEEE Trans. Syst. MAN, Cybern. B, vol. 30, no. 1, pp. 143–150, 2000.
[12] R. M. Murray, Z. Li, and S. S. Sastry, A Mathematical Introduction to Robotic Manipulation. CRC
Press, 1994.
[13] L. R. HUNT, R. SU, and G. MEYER, “Global Transformations of Nonlinear Systems,” IEEE
TRAMACTIONS Autom. Control, vol. AC-28, no. I, pp. 24–31, 1983.
[14] E. G. Gilbert and I. J. HA, “An Approach to Nonlinear Feedback Control with Applications to
Robotics,” IEEE Trans. ONSYSTEMS, MAN, Cybern., vol. SMc-14, no. 6, pp. 879–884, 1984.
[15] F. L. LEWIS, D. M. Dawson, and C. T. Abdallah, Manipulator Control Theory and Practice. NEW
YORK • BASEL: MARCEL DEKKER, INC., 2004.
[16] D. Receanu, “Modeling and Simulation of the Nonlinear Computed Torque Control in Simulink /
MATLAB for an Industrial Robot,” SL, vol. 10, no. 2, pp. 95–106, 2013.
[17] KUKA Roboter GmbH, Lightweight Robot 4+, Specification. D-86165 Augsburg, Germany, 2010.
ISSN: 2005-4238 IJAST 212

View publication stats

1212-ArticleText-1921-1-10-20191031

Uploaded by

Copyright:

Available Formats

1212-ArticleText-1921-1-10-20191031

Uploaded by

Document Information

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

1212-ArticleText-1921-1-10-20191031

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Dynamics and Computed-Torque Control of a 2-DOF manipulator:

Article in International Journal of Advanced Science and Technology · October 2019

Abdel-Nasser Sharkawy Panagiotis Koustoumpardis

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

Dynamics and Computed-Torque Control of a 2-DOF manipulator:

Abdel-Nasser Sharkawy1, 2*, Panagiotis N. Koustoumpardis2

Keywords:Robot-Manipulator, Joints Dynamics, Dynamic Coupling, Collision, Computed-Torque

ISSN: 2005-4238 IJAST 201

2. Dynamics of Multiple Joints Motion

In another form, the equation can be rewritten as[12]

𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝜏1

𝛼 = 𝐼𝑧1 + 𝐼𝑧2 + 𝑚1 𝑟12 + 𝑚2 𝐿21 + 𝑟22

ISSN: 2005-4238 IJAST 202

𝜏1 = 𝛼 + 2𝛽𝑐2 𝜃1 + 𝛿 + 𝛽𝑐2 𝜃2 − 𝛽𝑠2 𝜃1 𝜃2 − 𝛽𝑠2 (𝜃1 + 𝜃2 )𝜃2 (3)

𝜏2 = 𝛿 + 𝛽𝑐2 𝜃1 + 𝛿𝜃2 + 𝛽𝑠2 𝜃12 (4)

2.1. External Force is perpendicular to link 1 (𝐋𝟏 )

The work done by force 𝐹 is given by

𝑑𝑊 = 𝐹𝑟1 𝑑𝜃1 (5)

ISSN: 2005-4238 IJAST 203

The potential energy 𝑉 𝜃 is equal to the work done by force 𝐹 so

𝑉 𝜃 = 𝑑𝑊 = 𝐹𝑟1 𝑑𝜃1 (6)

𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝐹𝑟1 𝜏1

2.2. External Force is perpendicular to link 2 (𝐋𝟐 )

Figure 3. External perpendicular force (collision) exerted on link 2

The potential energy 𝑉 𝜃 is equal to the work done by force 𝐹

𝑉 𝜃 = 𝑑𝑊1 + 𝑑𝑊2 (9)

The work done by force 𝐹 at joint 2

𝑑𝑊1 = 𝐹𝑟2 𝑑(𝜃1 + 𝜃2 )(10)

ISSN: 2005-4238 IJAST 204

The work done by force F at joint 1

= 𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2 (11)

By substituting (10) and (11) into (9), the potential energy is

𝑉 𝜃 = 𝐹𝑟2 𝑑(𝜃1 + 𝜃2 ) + 𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2

= 2𝐹𝑟2 𝑑 𝜃1 + 𝜃2 + 𝐹𝐿1 cos 𝜃2 𝑑 𝜃1 + 𝜃2 (12)

Taking the derivative of (12) according to𝜃1 and 𝜃2

𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1

3. Computed Torque Control

ISSN: 2005-4238 IJAST 205

By taking the first and second derivative of the error is

From (1), 𝜃 is given by

By substituting 𝜃 from (17) into (16) so

Defining the control input function as [15]

From (19), the computed joint torque 𝜏 is given by

Select the control signal 𝑢 as the proportional-Derivative (PD) Feedback

ISSN: 2005-4238 IJAST 206

3.1. Computed Torque Control for Multiple joints motion

𝛼 + 2𝛽𝑐2 𝛿 + 𝛽𝑐2 𝜃1𝑑 + 𝐾𝑑1 𝑒1 + 𝐾𝑝1 𝑒1 −𝛽𝑠2 𝜃2 −𝛽𝑠2 (𝜃1 + 𝜃2 ) 𝜃1 𝜏𝑐1

A. If there is no external force

B. If there is external force that is perpendicular to link 1

ISSN: 2005-4238 IJAST 207

C. If there is external force that is perpendicular to link 2

4.1. Determining the Parameters

𝑓 is the frequency of the sinusoidal motion.

Table I. The values of the parameters used in the simulation[17]

ISSN: 2005-4238 IJAST 208

4.2. The Results

(a) Joint 1 (b) Joint 2

ISSN: 2005-4238 IJAST 209