International Journal of u- and e- Service, Science and Technology

Vol.8, No.1 (2015), pp.435-446

http://dx.doi.org/10.14257/ijunesst.2015.8.1.38

Active Disturbance Rejection Control of a SCARA Robot Arm

Ali Medjebouri1 and Lamine Mehennaoui2

1
Department of Mechanical Engineering, Skikda University, Algeria
2
Department of Electrical Engineering, Skikda University, Algeria
1
[email protected], [email protected]

Abstract
This paper is focused on a comparison between classic Feedback Linearization
Control and a proposed Active Disturbance Rejection Control method. The proposed
technique does not require a precise mathematical description of the system, since it is
based on the online estimation and rejection of the unmodeled elements of the dynamics.
Robustness of the closed-loop control system (against external perturbation and
parameters uncertainty) is discussed here. A model of a SCARA robot manipulator is used
in the conducted case study as an exemplary plant. Conclusions are supported with
results obtained with numerical simulations.

Keywords: SCARA, Position Control, Feedback Linearization Control, Active

Disturbance Rejection Control, Tracking Performances, Robustness

1. Introduction
The SCARA (selectively compliant assembly robot arm) robot arm, Figure 1, uses two
parallel revolute joints to produce motion in the horizontal plane. The arm structure is
weight-bearing but the first and second axes do no lifting. The third axis of the SCARA
provides work volume. By adding a fourth revolute joint, an additional rotation about the
z axis is added to control the orientation of the end effector in the horizontal plane. This
type of robots is rarely found with more than four axes. The SCARA robot manipulators
are used extensively in the assembly of electronic components and of small and medium
sized mechanical devices [1-3].
In controlling, the robot arm head is moved by actuators to attain an exact moving, and
the dynamic model of the robot manipulator is a MIMO (Multiple-Input Multiple-Output)
highly non-linear coupled system [2, 3].
Nonlinear systems control has been widely concern of the research. At present, the
nonlinear system decoupling control and static feedback linearization that based on the
theory of differential geometry brought the research getting rid of limitation for local
linearization and small scale motion [4]. However, differential geometry control must
depend on precise mathematical model, and when system works with various payloads
and has uncertainty in dynamic model this technique has limitations [5]. From control
point of view, uncertainty is divided into two main groups: uncertainty in unstructured
inputs (e.g., noise, external disturbance, and unmodeled dynamics) and uncertainty in
structure dynamics (parametric uncertainties). In some applications systems are used in an
unknown and unstructured environment, therefore strong mathematical tools are used in
new control methodologies to design nonlinear robust controllers with acceptable
performances (e.g., minimum error, good trajectory tracking, and disturbance rejection)
[6].
One ideal candidate is the Active Disturbance Rejection Control strategy (ADRC)
which has a very good disturbance rejection capability. The ADRC was firstly proposed
by Han in 1990s [7, 8] and further simplified and explicated by Gao in [9, 10]. The
essence of ADRC is that both the internal dynamics and the external disturbances can be

ISSN: 2005-4246 IJUNESST

Copyright ⓒ 2015 SERSC
International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

estimated and compensated in real time [6-8]. Due to its strong robustness and
disturbance rejection, ADRC has been successfully applied in many fields [7, 8, 11].
In this paper, we propose to improve classic feedback linearization control robustness
by introducing an active disturbance rejection component. The proposed strategy will be
applied for the position control of an electrically driven two–link SCARA robot arm.

(a) (b)

Figure 1.SCARA robot arm design:

(a). full manipulator.(b). End effector.

2. SCARA Horizontal Plane Dynamics Modeling

The SCARA robot uses the first two joints to move around in the horizontal plane.
Only these axes are dynamically coupled. The third and fourth axis can be considered as
decoupled mechanically from the motions of the other axes.
The manipulator powered by DC motors is modeled by the following dynamic
expression [1, 12]:
𝑀(𝑞). 𝑞̈ + 𝐶(𝑞, 𝑞̇ ) + 𝐹(𝑞, 𝑞̇ ) + 𝐺(𝑞) = 𝐾𝑇 𝐼 (1)
{
𝐿𝑎 𝐼 ̇ + 𝑅𝑎 𝐼 + 𝐾𝑒 𝑁𝑞̇ = 𝑉

Where symbols 𝑞, 𝑞̇ , and 𝑞̈ denote two-dimensional vectors of joint position, angular

velocity and angular acceleration. 𝑀(𝑞) ∈ 𝑅 2×2 is a symmetric, positive definite inertia
and mass matrix. 𝐶(𝑞, 𝑞̇ ) ∈ 𝑅 2×1 is the matrix of the Coriolis and centrifugal
forces. 𝐺(𝑞) ∈ 𝑅 2×1 is the vector of gravity terms, and 𝐹(𝑞, 𝑞̇ ) ∈ 𝑅 2×1 is the friction
torques vector considered as unknown external disturbances.
𝐼 ∈ 𝑅 2×1 is the armature current, 𝑉 ∈ 𝑅 2×1 is the armature voltage, and
𝐿𝑎 , 𝑅𝑎 , 𝐾𝑒 , 𝐾𝑇 , 𝑁 are positive definite diagonal matrices, representing respectively, the
actuators inductances, the actuators resistances, back electromotive force coefficients, the
constant coefficients characterizing the electromechanical conversion between current and
torque, and gearboxes ratios.
Since the first two axes only move the robot in the horizontal plane, the Gravity vector
is zero. The inertia, Coriolis and friction torques matrices for the proposed SCARA robot
arm are as follow [13]:

436 Copyright ⓒ 2015 SERSC

 International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

(𝐽1 + 𝐽2 ) 𝑚1 𝑟12 + 𝑚2 𝑟22 + 4𝑚2 𝑟12

𝑀11 = 𝐽𝑚1 + +
𝑁12 4𝑁12
𝑚2 𝑟1 𝑟2
+ (cos(𝑞1 + 𝑞2 ) cos(𝑞1 ) + sin(𝑞1 + 𝑞2 ) sin(𝑞1 ))
𝑁12

𝐽2 𝑚2 𝑟22 𝑚2 𝑟1 𝑟2
𝑀12 = + + (cos(𝑞1 + 𝑞2 ) cos(𝑞1 ) + sin(𝑞1 + 𝑞2 ) sin(𝑞1 ))
𝑁1 𝑁2 4𝑁1 𝑁2 2𝑁1 𝑁2
𝐽2 𝑚2 𝑟22 𝑚2 𝑟1 𝑟2
𝑀21 = + + (cos(𝑞1 + 𝑞2 ) cos(𝑞1 ) + sin(𝑞1 + 𝑞2 ) sin(𝑞1 ))
𝑁1 𝑁2 4𝑁1 𝑁2 2𝑁1 𝑁2
𝐽2 𝑚2 𝑟22
𝑀22 = 𝐽𝑚2 + +
𝑁22 4𝑁22
𝑚2 𝑟1 𝑟2 𝑞̇ 2 2
𝐶11 = (𝑞̇ 1 . 𝑞̇ 2 + ) (cos(𝑞1 + 𝑞2 ) sin(𝑞1 ) − sin(𝑞1 + 𝑞2 ) cos(𝑞1 ))
𝑁1 2
𝑚2 𝑟1 𝑟2 2
𝐶21 = (𝑞̇ 1 )(sin(𝑞1 + 𝑞2 ) cos(𝑞1 ) − cos(𝑞1 + 𝑞2 ) sin(𝑞1 ))
2𝑁2
𝐹11 = 𝑓𝑐1 𝑠𝑖𝑔𝑛(𝑞̇ 1 ) + 𝑓𝑣1 𝑞̇ 1
𝐹21 = 𝑓𝑐2 𝑠𝑖𝑔𝑛(𝑞̇ 2 ) + 𝑓𝑣2 𝑞̇ 2
In order to facilitate the control task, we propose to simplify (1) by neglecting the
armature inductance. Then current expression becomes:
(2)
𝐼 = 𝑅𝑎−1 . [𝑉 − 𝐾𝑒 𝑁𝑞̇ ]
Finally, system (1) is reduced to:
(3)
𝑀(𝑞). 𝑞̈ + ℎ(𝑞, 𝑞̇ ) = 𝐾. 𝑉
Where:
ℎ(𝑞, 𝑞̇ ) = 𝐶(𝑞, 𝑞̇ ) + 𝐹(𝑞, 𝑞̇ ) + 𝐾𝑇 . 𝑅𝑎−1 . 𝐾𝑒 𝑁𝑞̇
{
𝐾 = 𝐾𝑇 . 𝑅𝑎−1

3. Active Disturbance Rejection Control of the Robot Manipulator

Consider first the dynamic system given by the expression (3); the design of the
proposed controller is based on the following static nonlinear state feedback control,
where v is the new control vector:
𝐾. 𝑉 = 𝑀(𝑞). 𝑣 + ℎ(𝑞, 𝑞̇ ) (4)

If M is nonsingular, this control input simplifies (3) to the following linearized

decoupled system:
(5)
𝑞̈ = 𝑣
Taking into account the uncertainty of the system model, we can then express the
nonlinear control law (4) as follows:
̂ (𝑞). 𝑣 + ℎ̂(𝑞, 𝑞̇ ) (6)
𝐾. 𝑉 = 𝑀
̂ , ℎ̂ are the nominal values of M, h. The model errors are then represented as:
Where 𝑀

Copyright ⓒ 2015 SERSC 437

International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

∆𝑀 = 𝑀 ̂ (𝑞) − 𝑀(𝑞)
{ (7)
∆ℎ = ℎ̂(𝑞) − ℎ(𝑞)
From (1) and (6), we find:
̂ (𝑞). 𝑣 + ℎ̂(𝑞, 𝑞̇ ) = 𝑀(𝑞). 𝑞̈ + ℎ(𝑞, 𝑞̇ )
𝑀 (8)
This can be expressed as follows:
̂ . 𝑣 + 𝑀−1 . ∆ℎ
𝑞̈ = 𝑀−1 . 𝑀
̂ − 𝐼). 𝑣 + 𝑀−1 . ∆ℎ
= 𝑣 + (𝑀−1 . 𝑀
We then define:
̂ − 𝐼)
𝐸 = (𝑀−1 . 𝑀
{ (9)
𝜂 = 𝐸. 𝑣 + 𝑀−1 . ∆ℎ
Substituting (9) into (8), we obtain finally the effect of disturbances on the linearized
model:
𝑞̈ = 𝑣 + 𝜂(𝑣, 𝑞, 𝑞̇ ) (10)
In our case, the system dynamics are given by:
𝑞̈ = 𝑣1 + 𝜂1 (𝑣, 𝑞, 𝑞̇ )
{ 1 (11)
𝑞̈ 2 = 𝑣2 + 𝜂2 (𝑣, 𝑞, 𝑞̇ )
And the control outputs are:
𝑦1 = 𝑞1
{𝑦 = 𝑞 (12)
2 2
Both systems (11) can be rewritten as follows:
𝑥̇ 𝑖1 = 𝑥𝑖2
{ 𝑖2 = 𝑥𝑖3 + 𝑏. 𝑣𝑖
𝑥̇ (13)
𝑥̇ 𝑖3 = 𝜂𝑖̇ (. )
Where:
𝑥𝑖1 = 𝑞𝑖 |𝑖=1,2
𝑥𝑖2 = 𝑞̇ 𝑖 |𝑖=1,2
𝑥𝑖3 = 𝜂𝑖 (𝑣, 𝑞, 𝑞̇ )|𝑖=1,2
{ 𝑏=1
𝜂𝑖 (. ) is an unknown function which can be viewed as the total uncertainties or
disturbances of the system. Considering 𝜂𝑖 (. ) as an augmented state, it can be easily
observed. In [14, 16] we find the following nonlinear extended state observer (ESO:
Extended State Observer):
𝑥̂̇ = 𝐴𝑖 . 𝑥̂ + 𝐵𝑖 . 𝑣𝑖 + 𝐿𝑖 . 𝑓𝑎𝑙(𝑒𝑖 , 𝛼𝑖 , 𝛿𝑖 )
{ (14)
𝑦 = 𝐶𝑖 𝑥̂

438 Copyright ⓒ 2015 SERSC

 International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

Where:
0 1 0 0 𝐿𝑖1
𝐴𝑖 = [0 0 1] , 𝐵𝑖 = [𝑏], 𝐶𝑖 = [1 0 0], 𝐿𝑖 = [𝐿𝑖2 ]
0 0 0 0 𝐿𝑖3
The observer errors are:
𝑒𝑖 |𝑖=1,2 = 𝑞𝑖 − 𝑥̂1 (15)
Functions 𝑓𝑎𝑙𝑖𝑗 are defined as:
|𝑒𝑖 |𝛼𝑖𝑗 . 𝑠𝑖𝑔𝑛(𝑒𝑖 ), 𝑖𝑓 |𝑒𝑖 | > 𝛿𝑖
𝑓𝑎𝑙𝑖𝑗 (𝑒𝑖 , 𝛼𝑖𝑗 , 𝛿𝑖 )|𝑖=1,2 = { 𝑒𝑖 , (16)
1−𝛼𝑖𝑗 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑗=1,2,3 𝛿𝑖

Here 0 < 𝛼𝑖𝑗 < 1 is the nonlinear coefficient, 𝛿𝑖 > 0 is the turning point of fal
function. Note that, if the error is small (|𝑒𝑖 | ≤ 𝛿𝑖𝑗 ); the proposed ESO becomes a simple
linear observer (LESO):
𝑥̇ 𝑖1 = 𝑥𝑖2 + 𝛽𝑖1 𝑒𝑖
{𝑥̇ 𝑖2 = 𝑥𝑖3 + 𝛽𝑖2 + 𝑏𝑣𝑖 (17)
𝑥̇ 𝑖3 = 𝛽𝑖3 𝑒𝑖
Where:
𝐿𝑖𝑗
𝛽𝑖𝑗 = 1−𝛼𝑖𝑗 |
𝛿𝑖 𝑖=1,2
𝑗=1,2,3

The parameters 𝛽𝑖𝑗 are chosen in a special way as [7-9, 14]:

𝑠 3 + 𝛽𝑖1 𝑠 2 + 𝛽𝑖2 𝑠 + 𝛽𝑖3 = (𝑠 + 𝜔0𝑖 )3 |𝑖=1,2 (18)
Where 𝜔0𝑖 denotes the bandwidth of the LESO (17).
Once uncertainties and disturbances effects are totally estimated by the ESO, we can
now compensate them in real time by the use of the control signal vi [8-10, 14, 15, 17]:
(𝑣𝑖0 − 𝑥̂𝑖3 )⁄
𝑣𝑖 = 𝑏
(19)

Appliying this control input, system (11) is reduced to:

𝑞̈ 𝑖 = (𝑥𝑖3 − 𝜂𝑖 ) + 𝑣𝑖0 (20)
When 𝑥̂𝑖3 → 𝜂𝑖 :
𝑞̈ 𝑖 ≈ 𝑣𝑖0 (21)
The expression (21) represents a double integrator with unit gain, which can be readily
driven to the desired position by the use of a simple linear controller. Let 𝑞𝑑𝑖 denote the
reference position for each joint, and the tracking errors 𝑒𝑖 = 𝑞𝑑𝑖 − 𝑞𝑖 . Then the linear
controller can be defined as:
2
𝑣𝑖0 = 𝑞̈ 𝑑𝑖 + 2𝜁𝑐𝑖 𝜔𝑐𝑖 𝑒𝑖̇ + 𝜔𝑐𝑖 𝑒𝑖 (22)
Which leads to the following second order tracking dynamics in the closed loop:
2
𝑒̈𝑖 + 2𝜁𝑐𝑖 𝜔𝑐𝑖 𝑒𝑖̇ + 𝜔𝑐𝑖 𝑒𝑖 = 0 (23)

Copyright ⓒ 2015 SERSC 439

International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

Where 𝜁𝑐𝑖 , 𝜔𝑐𝑖 are respectively the damping ratio and the natural frequency of the
closed-loop system. The bandwidth 𝜔0𝑖 of the ESO is chosen to ensure a faster dynamics
than (23):
𝜔0𝑖 = (3 𝑡𝑜 5)𝜔𝑐𝑖 (24)
The block diagram of the ADRC strategy is shown in Figure 2.
𝑤: Disturbance

Linear + 𝑣 System 𝑦
𝑦𝑑 1/b
Controller 𝑣0 −

𝑥̂1 𝑥̂2 𝑥̂3 +

Noise
+
ESO

Figure 2. Block Diagram of the ADRC Control Strategy

Note that if the active rejection component 𝑥̂𝑖3 is not introduced in the control law (19),
the proposed technique is reduced to the classical feedback linearization control method
(FLC).

4. Results and Discussion

The above proposed control strategy for a two–link SCARA robot manipulator
electrically controlled has been verified through computer simulations using
Matlab/Simulink. The block diagram of the proposed control strategy is shown in the
Figure 3, and the parameters of the considered robot manipulator are those indicated in
the appendix.
For measuring the performances in both tracking (reaching time, accuracy) and
regulation (disturbances rejection) modes we will consider two cases:

Case1(tracking performances measurement): the control law is applied on the

nominal system withaout considering internal or external disturbaces effects
(parameters uncertaintys and friction torques respectively). Since the active
rejection terme is inactivated, the proposed control method is reduced to the
classical feedback lineariszation control scheme (FLC).

Case2 (disturbances rejection measurement): friction torques considered as

external disturbances and a wide parameters uncertainty ∆𝑀 = 50%𝑀 ̂ (𝑞), ∆ℎ =
̂
50%ℎ(𝑞) are introduced to the manipulator model, and the proposed technique
with an active disturbance rejection control component (ADRC) is compared to the
classical feedback linearization control strategy (FLC).

440 Copyright ⓒ 2015 SERSC

 International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

x 1e. x1e
Linear controller 1
x 2e.
q1d x 3e. x2e
v 01
q1
Sine Wave u
x3e
y
Subtract 1
ESO1

q1 q1
q2 V1 V1 Scope 2
q2
q1_p

Linear Controller 2 q2_p q1_p

q2d u1
V2 V2
v 02 q2_p
q2
Sine Wave 1 u2

Subtract Non _linear _Control SCARA +Actuators

x 1e. x1e
x 2e.
x 3e. x2e
u
x3e
y
ESO2

Figure 3. Block Diagram of the Proposed Control Strategy

The obtained results by the 4th order Runge Kutta solver with a step size ∆t=0.001sec
are given below:

Case1:
Link 1 Link 2
reference reference
1.5 ADRC 1.5 ADRC
angular position of link1 [rad]

angular position of link2 [rad]

FLC FLC
1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
0 5 10 0 5 10
time(sec) time(sec)

0.6 1
tracking error of link1 [rad]

tracking error of link2 [rad]

ADRC ADRC
0.4
FLC FLC
0.2 0.5

-0.2 0

-0.4

-0.5
0 5 10 0 5 10
time(sec) time(sec)

Figure 4. Tracking Performances

Copyright ⓒ 2015 SERSC 441

International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

Link 1 Link 2
1.5

control signal of link1 [V]

control signal of link1 [V]

6 ADRC FLC
FLC 1 ADRC
4
0.5
2
0

0
-0.5
0 5 10 0 5 10
time(sec) time(sec)

Figure 5. Control inputs

Link 1 Link 2

x11 x21
x11est x21est
estimate of x11[rad]

estimate of x21[rad]

1 1

0 0

-1 -1

0 5 10 0 5 10
time(sec) time(sec)
estimate of x12[rad/sec]

estimate of x22[rad/sec]

4 x12 4 x22
x12est x22est
2 2

0 0

-2 -2
0 5 10 0 5 10
time(sec) time(sec)
estimate of x23[rad/sec2]
estimate of x13[rad/sec2]

100
x13 100 x23
x13est x23est
0 0

-100
-100
0 5 10 0 5 10
time(sec) time(sec)

Figure 6. ESO Estimated States Without Disturbances

442 Copyright ⓒ 2015 SERSC

 International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

0.15 0.4
estimate of x13[rad/sec2]

estimate of x23[rad/sec2]
0.1
0.2
0.05
0 0

-0.05 -0.2
-0.1
-0.4
-0.15
-0.6
4.13 4.135 4.14 4.145 2 2.01 2.02 2.03
time(sec) time(sec)

Figure 7. Zoom on x13 and x23 Estimated States

Case 2:
Link 1 Link 2
reference reference
ADRC ADRC
1.5 1.5
FLC FLC
angular position of link1 [rad]

angular position of link2 [rad]

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
0 5 10 0 5 10
time(sec) time(sec)
ADRC ADRC
FLC FLC
0.6 1

0.4
tracking error of link1 [rad]

tracking error of link2 [rad]

0.5
0.2

0 0

-0.2
-0.5
-0.4

-1
0 5 10 0 5 10
time(sec) time(sec)

Figure 8. Tracking performances in Presence of Disturbances

Copyright ⓒ 2015 SERSC 443

International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

Link 1 Link 2
7 1.5
FLC FLC
6 ADRC ADRC

control signal of link1 [V]

control signal of link2 [V]

5 1

4
0.5
3

2
0
1

0
-0.5
0 5 10 0 5 10
time(sec) time(sec)

Figure 9. Control Inputs in Presence of Disturbances

Link1 Link2
x11 x21
x11est x21est
estimate of x11[rad]

estimate of x21[rad]

1 1

0 0

-1 -1

0 5 10 0 5 10
time(sec) time(sec)
x22
x12
estimate of x12[rad/sec]

estimate of x22[rad/sec]

4 4 x22est
x12est
2 2

0 0

-2 -2

-4 -4
0 5 10 0 5 10
time(sec) time(sec)
x13 x23
estimate of x13[rad/sec2]

estimate of x23[rad/sec2]

100 x13est x23est

100
50

0 0

-50
-100
-100
0 5 10 0 5 10
time(sec) time(sec)
Figure 10. ESO Estimated States in Presence of Disturbances

444 Copyright ⓒ 2015 SERSC

 International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

Case1: Since the system is not subject to disturbances effects, and as provided both
strategys give exactly the same good results in tracking mode, Figure 4. The Figure 6
shows clearly the convergence of the observed states to the real system states.
Case2: The simulation results presented in the Figure 8 highlight the effectiveness
and the superiority of the proposed ADRC control strategy compared with classical
feedback control law (FLC) in presence of wide parametric and external
disturbances. This means that the extended state observer was able to observe and
decrease disturbances effects which is confirmed in Figure 10. The Figure 9 shows
the control inputs and the compensation components effects.

5. Conclusion
In this paper, active disturbance rejection control method (ADRC) was successfully
applied to the position control of a two-link SCARA robot arm electrically actuated.
Simulation results demonstrate that the proposed control algorithm can provide good
tracking performance and well handle disturbances for the robot manipulator.
It should be noted that the design of this control strategy does not require a high
accuracy mathematical model of the controlled system. This practical control strategy is
easy to understand and implement, making it an appealing method to real applications.

Appendix:
Parameters of the robot arm:
𝑚1 = 1,90 kg, 𝑚2 = 0,93 kg, 𝐽1 = 0,0980 kgm2 , 𝐽2 = 0,0115kgm2 , 𝐽𝑚1 = 𝐽𝑚2 =
3,3. 10−6 kgm2 , 𝑟1 = 0,25𝑚, 𝑟2 = 0,15𝑚, 𝑅𝑎1 = 𝑅𝑎2 = 3,5 Ω, 𝐿𝑎1 = 𝐿𝑎2 = 1,3 mH,
𝐾𝑒1 = 𝐾𝑒2 = 0,047 V/rad/s, 𝐾𝑇1 = 𝐾𝑇2 = 0,047 Nm/A, 𝑁1 = 90, 𝑁2 = 220.
Parameters of the extended state observer:
̅̅̅̅ 𝑎𝑛𝑑 𝑗 = 1,3
𝛼𝑖𝑗 = 0,5; 𝛿𝑖 = 0,01; 𝜔0𝑖 = 5𝜔𝑐𝑖 𝑓𝑜𝑟 𝑖 = 1,2 ̅̅̅̅ .
Parameters of the linear controller:
̅̅̅̅
𝜁𝑐𝑖 = 1; 𝜔𝑐𝑖 = 10 𝑟𝑎𝑑/𝑠𝑒𝑐 𝑓𝑜𝑟 𝑖 = 1,2

References
[1] N. Munro and F. L. Lewis, Robot Manipulator Control-Theory and Practice, 2nd edition, Marcel Dekker
Inc., (2004).
[2] A. A. Amiri-M, M. R. Gharib, M. Moavenian and K. Torabi Z. Modeling and control of a SCARA robot
using quantitative feedback theory. Journal of Systems and Control Engineering. Part I, 223 (2009).
[3] K. Sahari, K. H. Weng, Y. W. Han, A. Anuar, M. Z. Baharuddin and S. Mohideen. Design and
Development of a 4-DOF SCARA Robot for Educational Purposes. Jurnal Teknologi, Universiti
Teknologi Malaysia. (2011).
[4] C. Kravaris and C. B. Chung. Nonlinear State Feedback Synthesis by Global Input /Output Linearization.
AIChE Journal. 4, 33 (1987).
[5] C. Kravaris and S. Palanki. Robust Nonlinear State Feedback under Structured Uncertainty. AIChE
Journal. 7, 34 (1988).
[6] F. Piltan, M. R. Rashidian, M. Shamsodini and S. Allahdadi. Effect of Rule Base on the Fuzzy-Based
Tuning Fuzzy Sliding Mode Controller: Applied to 2nd Order Nonlinear System. International Journal
of Advanced Science and Technology. 46 (2012).
[7] S. Liu, X. Mei and F. Kong. Research on Unwinding Tension System Control Based on Active
Disturbance Rejection Control. 2nd International Conference on Materials, Mechatronics and
Automation, (2012) May 7-8, 2012, Nanchang, China, pp. 252-257.
[8] J. Zhang, J. Feng, Y. Zhou, F. Fang and H. Yue. Linear Active Disturbance Rejection Control of Waste
Heat Recovery Systems with Organic Rankine Cycles. Energies. 5 (2012).
[9] Z. Gao. Active Disturbance Rejection Control: A Paradigm Shift in Feedback Control System Design.
Proceedings of the 2006 American Control Conference, (2006) June 14-16; Minnesota, USA, pp. 2399-
2405.

Copyright ⓒ 2015 SERSC 445

International Journal of u- and e- Service, Science and Technology
Vol.8, No.1 (2015)

[10] Q. Zheng, L. Q. Gao and Z. Gao. On Stability Analysis of Active Disturbance Rejection Control for
Nonlinear Time-Varying Plants with Unknown Dynamics. Proceedings of the 46th IEEE Conference on
Decision and Control. (2007) December 12-14; New Orleans, USA, pp. 3501-3506.
[11] Q. Zheng and Z. Gao. On Practical Applications of Active Disturbance Rejection Control. Proceedings
of the 29th Chinese Control Conference. (2010) July 29-31; Beijing, China, pp. 6095-610.
[12] N. Azoui and L. Saidi. Passivity Based Adaptive Control of Robotic Manipulators Electrically
Controlled. International Journal of Advanced Science and Technology. 34 (2011).
[13] M. T. Das and L. C. Dülger. Mathematical Modeling, Simulation and Experimental Verification of a
SCARA Robot. Simulation Modelling Practice and Theory. 13 (2005).
[14] J. Han. From PID to Active Disturbance Rejection Control. IEEE Transactions on Industrial Electronics.
3, 56 (2009).
[15] Q. Zheng, L. Q. Gao and Z. Gao. On Estimation of Plant Dynamics and Disturbance from Input-Output
Data in Real Time. 16th IEEE International Conference on Control Applications. (2007) October 1-3;
Singapore, pp.1167-1172.
[16] X. Yang and Y. Huang. Capabilities of Extended State Observer for Estimating Uncertainties. 2009
American Control Conference. (2009) June 10-12; St. Louis, USA, pp. 3700-3705.
[17] A. Boukhriss, T. Nasser and A. Essadki. A Linear Active Disturbance Rejection Control applied for
DFIG based Wind Energy Conversion System. International Journal of Computer Science Issues. 2, 10
(2013)

Authors
Ali Medjebouri was born in Skikda, Algeria in 1984. He received
his Engineer degree in Automatic Control Systems in 2007 and his
magister degree in Mechatronics in 2010 from the University of 20
August 1955, Skikda, Algeria. Currently, he is a PhD student. His
research interests concern robotics and system’s control.

Lamine Mehennaoui was born in Constantine,Algeria, in 1967.

He received the Engineer Degree in 1990, the Magister. Degree in
1994 and the PhD. Degree in 2006 all in Automatic Control Systems
from the University of Baji Mokhtar, Annaba, Algeria. Since 2003,
he was an affiliated researcher in the laboratory Automatic Skikda.
Currently he is assistant professor in the Electrical Engineering
Department of the University of 20 August 1955, Skikda Algeria. His
research interests concern robotics and system’s control.

446 Copyright ⓒ 2015 SERSC