International Journal of Advanced Engineering Research and

Science (IJAERS)
Peer-Reviewed Journal
ISSN: 2349-6495(P) | 2456-1908(O)
Vol-10, Issue-8; Aug, 2023
Journal Home Page Available: https://ijaers.com/
Article DOI:https://dx.doi.org/10.22161/ijaers.108.6

A Fuzzy Robust Controller for Robotic Systems

Lafi Alnufaie

Department of Electrical Engineering, Engineering College, Shaqra University, Shaqra 11961, Saudi Arabia

Received: 11 Jul 2023, Abstract— In this work a combination of type 2 fuzzy logic and
Receive in revised form: 14 Aug 2023, nonsingular fast sliding mode technique is proposed to design a robust
controller for a robotic system. Indeed, a nominal type 2 fuzzy model is
Accepted: 22 Aug 2023,
used to contract the equivalent control signal. The switching signal is
Available online: 31 Aug 2023 designed using adaptive type 2 fuzzy systems to overcome the knowledge of
©2023 The Author(s). Published by AI the upper bounds of uncertainties and external disturbances. Several
Publication. This is an open access article simulation results are given to show the efficiency of the proposed
under the CC BY license approach.
(https://creativecommons.org/licenses/by/4.0/).
Keywords— Type-2 Fuzzy Systems,
Robust Control, Robotic system.

I. INTRODUCTION powers. This problem can be resolved by using a

Widely used in many applications, sliding mode control nonsingular terminal sliding mode controller [7-8].
can be considered a very popular approach to ensure good Nevertheless, this improvement was obtained at the
tracking performances against external disturbances [1]. expense of the convergence time which becomes slower.
Despite its simple design procedure and good tracking Nonsingular fast terminal sliding mode controller have
performances, sliding mode control has two major been developed to overcome singularity and to obtain fast
disadvantages. Thes first one is the chattering phenomena convergence time [9]. Thus, in this paper, we propose a
introduced by using signum function in the control signal. nonsingular fast terminal sliding mode controller for a
The second disadvantage lies in its time convergence, robotic system which guarantees finite-time convergence,
which cannot impose. Several improvements have been fast speed when the states are far from the origin,
proposed in the literature to reduce the chattering avoidance of singularity and without chattering. The
phenomena. In [2], the switching signal is smoothed by control is developed using a fuzzy nominal model witch
using a low-pass filter. An adaptive fuzzy system has been avoids using approximating system dynamics.
used in [3] to substitute the switching control and, hence, Furthermore, adaptive type 2 fuzzy systems have been
to eliminate the chattering phenomenon. However, this used to avoid a well-knowledge of the upper bounds of
improvement needs a tradeoff between the smoothness of both uncertainties and external disturbances.
the switching signal and tracking performances. Second The remainder of this paper is organized as follows:
order sliding mode control have been also presented a Section 2 is dedicated to introduce type 2 fuzzy systems.
good solution to chattering but the design procedure is In Section 3, problem statement of controlling a robotic
complex and the requires a good knowledge of the studied system is treated. Section 4 is dedicated to the controller
system [4]. design and stability analysis. Simulation and results are
Recently, terminal sliding mode control have been given in Section 5 to show the effectiveness of the
developed, where a nonlinear surface is used [5-6]. proposed approach. Finally, the conclusion is provided.
However, these kinds of controllers suffer from singularity
problem due to presence of terms with negative fractional

www.ijaers.com Page | 45
Alnufaie et al. International Journal of Advanced Engineering Research and Science, 10(8)-2023

II. INTERVAL TYPE-2 FUZZY LOGIC SYSTEM Where:

Fuzzy Logic Systems are known as the universal
approximators and have several applications in control
design and identification. A type-1 fuzzy system consists
of four major parts: fuzzifier, rule base, inference engine,
In this work, the center of set type-reduction method is
and defuzzifier. A T2FLS is very similar to a T1FLS [10]
used to simplify the notation. Therefore, the output can be
the major structure difference being that the defuzzifier
expressed as:
block of a type-1 fuzzy system is replaced by the output
processing block in a type-2 fuzzy system, which consists
of type-reduction followed by defuzzification. Where is also an interval type 1 set determined by
left and right most points ( and , witch can be derived
from consequent centroid set (either or ) and

the firing strength . The interval set

( ) should be computed or set first
before the computation of . Hence, left most point
Fig. 1: Structure of a type-2 fuzzy logic system
and right most point can be expressed as:

In a interval type-2 fuzzy system, a triangular fuzzy set is

defined by a lower and upper set as shown in figure 2. (1)

Using the center of set type reduction method to compute

and the defuzzified crisp output from an interval type
2 fuzzy logic system can be obtained according to the
following equation:
(2)

Which can be rewritten on the following vectorial form:

(3)
Fig. 2: Interval type-2 triangular fuzzy sets
It is clear that the interval type-2 fuzzy set is in a region Where represents the regressive vector and the
bounded by an upper membership function and a lower consequent vector containing the conclusion values of the
membership function denoted as and respectively fuzzy rules.

and is named a foot of uncertainty (FOU). Assume that

there are M rules in a type-2 fuzzy rule base, each of them III. PROBLEM STATEMENT
has the following form: Let us consider the dynamic equation of n degree-of-
freedom robotic manipulators as follows

Where and are the input and output

variables of variables of the type 2 fuzzy system, (4)
respectively, the is the type 2 fuzzy sets of antecedent where , and are the vector of joint position,
part, and is the weighting interval set in the joint velocity, and joint acceleration, respectively.
is a symmetric and positive definite inertia
consequent part. The operation of type-reduction is to give
a type-1 set from a type-2 set. In the meantime, the firing matrix, is the matrix of centrifugal and
strength Fi for the ith rule can be an interval type-2 set Coriolis forces, is the vector of gravitational
expressed as: forces, is the vector of input joint torque and
is the vector of unknown external
disturbances.

www.ijaers.com Page | 46
Alnufaie et al. International Journal of Advanced Engineering Research and Science, 10(8)-2023

For practical applications, it is impossible to know the The time derivative of the sliding surface can be written
exact dynamic model of the robotic manipulators. Hence, as:
the above dynamic quantities can be expressed as: (10)
Our control law will be composed from two terms. The
(5)
first one, named equivalent control , is dedicated to
maintain the system on the sliding surface. The second
Where , , are the nominal values of term, called switching signal, must force the system
, , respectively and , , to converge to the sliding surface. Then, to design the
are the uncertain parts of , , equivalent control law , we consider that the system
respectively. is on the surface ( ) and remains on ( ). In
Using equation (5), the dynamic model of the robotic this case, the system is considered insensitive to
manipulators can be expressed as: uncertainties and external disturbances (Slotine, 1991).
Using (7) equation (10) can be rewritten as:
(6)
Where:
. (11)
Then the expression of equivalent control law can be
Let define the tracking error and its time
expressed as:
derivative where the desired trajectory.
Then the error dynamic of the robotic manipulators with
the uncertainties and disturbances can be written as:
(12)
(7)
Note that, we used the fact that to writ
Where ,
equation (9) in a compact form.
and
Our next task is to determine the expression of the
As given in [14], the upper bound of lumped uncertainty switching signal allowing to force the system to
can be expressed as:
reach the sliding surface in presence of uncertainties and
(8) external disturbances.
Where , and are positive scalars. In this case, equation (10) becomes:
The next task is to develop a robust controller based on
nonsingular fast terminal sliding mode control allowing to
tracking objectives. (13)
Using (12), we can rewrite (10) as:
IV. CONTROLLER DESIGN
To design our controller, let consider the following
nonsingular terminal sliding surface: (14)

(9) According to the definition of the equivalent control,

equation (14) can be simplified to:
Where and are positive constants, and
(15)
.
To deduce the expression of allowing the switching
The structure of this surface allows us to attain fast
convergence of the tracking error to zero. Indeed, if the condition, we consider the following Lyapunov function:
position initial value is far from the desired one, then the (16)
term will be dominant, which leads to a
fast convergence. In the case where the system is near the Differentiating with respect to time and using (15)
desired trajectory, the term must ensuring lead to:
a finite time convergence. (17)

www.ijaers.com Page | 47
Alnufaie et al. International Journal of Advanced Engineering Research and Science, 10(8)-2023

Choosing as: Nevertheless, it is very difficult if not possible to know the

exact values of the scalars , and . To overcome this
problem, we propose to approximate them by three
(18) adaptive type 2 fuzzy systems ,
and . According to
Where and are two positive scalars.
the universal approximation theorem, there exists an
The time derivative of the Lyapunov function becomes: optimal values of type 2 fuzzy systems we can write:

(26)

Where , and represent the optimal values of ,

and respectively.
(19)
Consequently, the control laws become:
Using the assumption (8), we obtain the following
inequality:

(20) (27)
Based on the Lyapunov theorem, the system converges These modified control laws allow to ensure the
asymptotically to the sliding surface and remains on. convergence to the reference trajectory in a finite time.
To prove convergence in finite time, let us take up To deduce the adaptation laws of the three adaptive fuzzy
inequality (20): system, we consider the new Lyapunov function:

(21)

(28)
Using the control laws (27) and the following adaptation
laws:
(22)
Then we can obtain: (29)

(23)
And following the same mathematical development used
previously, the time derivative of the Lyapunov function
(28) becomes:
If we consider that the system converges to 0 at
implies that:
(30)
Thus, the convergence of the closed loop system to the
reference trajectory in a finite time is guaranteed.
(24)
Hence, V. SIMULATION AND RESULTS
To show the performances of the performances of the
(25)
proposed approach, we consider a one link robot, shown in
fig. 3, whose dynamics equation is given by:
Consequently, the control law ,
whose terms are defined by equations (12) and (18),
guarantees the asymptotic stability of the closed loop With:
system and the convergence of the tracking error in a finite
time.

www.ijaers.com Page | 48
Alnufaie et al. International Journal of Advanced Engineering Research and Science, 10(8)-2023

Fig. 3: One link robot manipulator

To construct the type 2 fuzzy nominal model, we consider

that the position is constrained within , which
leads to 3 fuzzy rules. Each one of them gives the relation
between the equilibrium point and the corresponding local
model. Then, each rule uses a type 2 fuzzy sets in the
antecedent part to describe the equilibrium point and the Fig. 4: Angular position and velocity tracking
consequent part the corresponding local model. Using the
product as an interference engine, the method of center set
for the reduction type and center of gravity for
defuzzification, the output fuzzy system will be giving the
type 2 fuzzy nominal model.
Fig. 4 gives the angular position and velocity for two
initial positions. The convergence to zero in a finite time is
well shown in fig. 5. To illustrate the efficiency of the
proposed approach we have used a more complex
trajectory,
. Fig. 6 gives the angular
position and velocity for two initial values. These results
show also the good tracking performances and the
convergence to the reference trajectory in a finite time.
Furthermore, the control signal given by fig. 7 the
elimination of the chattering phenomenon and the smooth
control signal. We can conclude that the proposed Fig. 5: Error phase plane
approach ensures high tracking precision, fast response,
singularity avoidance and strong robustness to external
disturbances and modeling uncertainties.

www.ijaers.com Page | 49
Alnufaie et al. International Journal of Advanced Engineering Research and Science, 10(8)-2023

REFERENCES
[1] J.-J. E. Slotine, W. Li, and others, Applied nonlinear
control, vol. 199. Prentice hall Englewood Cliffs, NJ, 1991.
[2] Y. Pan, C. Yang, L. Pan and H. Yu, "Integral Sliding Mode
Control: Performance, Modification, and Improvement," in
IEEE Transactions on Industrial Informatics, vol. 14, no. 7,
pp. 3087-3096, July 2018.
[3] A. Hamzaoui, N. Essounbouli, and J. Zaytoon, “Fuzzy
sliding mode control with a fuzzy switching function for
non-linear uncertain multi-input multi-output systems,”
Proc. Inst. Mech. Eng. Part J. Syst. Control Eng., vol. 218,
no. 4, pp. 287–297, 2004.
[4] M. Manceur, L. Menhour, N. Essounbouli, and A.
Hamzaoui, “MIMO sliding fuzzy type-2 control with
manipulating approaching phase,” in 2013 10th IEEE
International Conference on Networking, Sensing And
Control (ICNSC), IEEE, 2013, pp. 479–485.
[5] P. Vivekanandan et al., “Modelling, Simulation and Control
of 5 Axis Industrial Robot using MATLAB,” Indian J. Sci.
Technol., vol. 12, no. 23, pp. 1–7, Jun. 2019.
[6] H. Wang et al., “Design and implementation of adaptive
terminal sliding-mode control on a steer-by-wire equipped
Fig. 6: Angular position et velocity tracking road vehicle,” IEEE Trans. Ind. Electron., vol. 63, no. 9, pp.
5774–5785, 2016.
[7] Y. Feng, X. Yu and Z. Man, "Non-singular terminal sliding
mode control and its application for robot
manipulators," ISCAS 2001. The 2001 IEEE International
Symposium on Circuits and Systems (Cat. No.01CH37196),
Sydney, NSW, Australia, 2001
[8] Alattas, K.A.; Mostafaee, J.; Alanazi, A.K.; Mobayen, S.;
Vu, M.T.; Zhilenkov, A.; Abo-Dief, H.M. Nonsingular
Terminal Sliding Mode Control Based on Adaptive Barrier
Function for nth-Order Perturbed Nonlinear
Systems. Mathematics 2022,
[9] Wang, L. and Yang, J., Nonsingular fast terminal sliding-
mode control for nonlinear dynamical systems. Int. J.
Robust Nonlinear Control, 21: 1865-1879, 2011.
[10] N. N. Karnik, J. M. Mendel, and Q. Liang, “Type-2 fuzzy
logic systems,” IEEE Trans. Fuzzy Syst., vol. 7, no. 6, pp.
643–658, 1999.

Fig. 7: Applied control signal

VI. CONCLUSION
In this paper, a combination of type 2 fuzzy logic and
nonsingular fast sliding mode control for robotic systems
is presented. The proposed approach allows to overcome
the drawbacks encountered in classical cases, thanks to
type 2 fuzzy logic. Several simulation results have
presented to show the efficiency of the proposed approach
despite the presence of modeling uncertainties and external
disturbances. As perspective of this work to improve the
control law design by reducing number of used
parameters.

www.ijaers.com Page | 50