Professional Documents
Culture Documents
Thanh-Tung Bui
Department of Applied Sciences
International School, Vietnam National University
Hanoi, Vietnam
buithanhtung@vnu.edu.vn
Abstract—This paper is concerned with a novel adaptive fault component of the control law is needed to compensate for
tolerant control (FTC) method for a wheeled mobile robot the deviations between the actual and estimated values of
under actuator fault and dead zone to track a predefined the magnitudes of the actuator faults, which is the reason
trajectory. Prior knowledge of actuator fault and a dead zone
is not needed. Although belonging to the passive category, the why the large-magnitude chattering phenomenon occurs [5]. In
proposed FTC method does not require a robust component that the active trend, the fault information is obtained via a fault
causes a chattering phenomenon to overcome the actuator fault. observer also known as fault diagnosis mechanisms. Hence,
Instead, the actuator fault is dealt with by a unique technique. not only the state of the control system is monitored carefully
The convergence of tracking errors is ensured by Lyapunov for maintenance but also the fault information is provided to
criterion. The merit and effectiveness of the proposed method
are confirmed via simulation results. the controller for compensating for the actuator faults [6].
Index Terms—dead-zone, fault-tolerant control, track trajec- Nonetheless, the accuracy of the observers strongly affects
tory, wheeled mobile robot. the performance of the control systems in the active trend.
If the observers, particularly, supply wrong estimations, the
I. I NTRODUCTION performance plummets, whereas heightening the reliability of
It is well-known that wheeled mobile robots (WMRs) have the observers, in contrast, makes the cost of the fault-tolerant
been applied all over the world. Their applicability makes control system remarkably surge [7].
applications more flexible such as e-commerce, warehouses, It is vital to consider the dead zone issue which is a non-
search and rescue, reconnaissance, and so forth. Many works sensitive phenomenon on the inputs of actuators [8] inside
studying to control a WMR for tracking a trajectory under the actuators. The performance of control systems rapidly
nonholonomic constraints have been implemented. deteriorates unless such a dead-zone issue is dealt with [9].
During the operation, there might be faults happening at Motivated by the aforementioned studies, we propose a
actuators [1]. Hence, the task of ensuring the reliability and novel adaptive fault-tolerant control method for a wheeled
safe of the actuators is vital. With the aim of minimizing the mobile robot to track the position of a reference trajectory
influence of actuator faults, there are so many solutions of under the actuator fault and the dead zone. The highlight
fault-tolerant control for objects under such actuator faults contribution of our proposed work is that even though our
[2]. It is emphasized that there have been two major trends fault-tolerant control solution belongs to the passive trend, it
to research fault-tolerant control, namely passive and active still does not need a large-magnitude robust control component
trends. In the passive trend, adaptive laws are utilized in order to overcome the actuator fault. This advantage is similar to the
to update the states of the actuators and then the controller approaches belonging to the active fault-tolerant control trend
will tackle such states [3]. However, the shortcoming of this rather than the passive one. Furthermore, this work continues
trend is that the exact magnitude of the actuator faults is the research of [7] by considering the saturation issue.
tough to estimate [4]. Consequently, a large-magnitude robust
The remainder is organized as follows. Section 2 shows the
*This research is funded by International School, Vietnam National Univer- background, followed by Section 3 illustrating the main result.
sity, Hanoi (VNU-IS) under project number CS.2022-01. Next, the simulation results are displayed in Section 4. Finally,
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
Page 158
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
Page 159
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
Page 160
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
For the evolution of both ρ̂ and κ̂, the following adaptive − sT H Q̂ diag sign(u) min u, γ − u
laws are proposed to ensure the stability of the entire control
system: + sT H Q̂ diag sign(v) min v, γ − v
ρ̂˙ = αH Q̂diag s u,
(26) + sT diag ρ̃ u − ρ̃T diag s u
˙κ̂ = −βH Q̂diag s sign(χ).
− sT diag κ̃ sign χ + κ̃T diag s sign χ .
(27) (33)
where α and β are positive constants selected arbitrarily. Furthermore, it is undoubted that the three following equa-
If ∥ψ∥ ≥ ϵ with ϵ being a positive small constant, then tions and one inequality hold:
sT diag ρ̃ = ρ̃T diag s ,
ψ̇ = − Γψ − diag(s)ψ (34)
sT diag κ̃ = κ̃T diag s ,
ψ (35)
sT H Q̂ diag sign(u) min u, γ − u
− 2
(28)
∥ψ∥ sT diag ψ = ψ T diag s ,
(36)
where Γ is a positive-definite constant matrix.
diag sign(u) min u, γ − u ≥
Otherwise,
diag sign(v) min v, γ − v . (37)
ψ̇ = 0
Substituting the equations and inequality into (32) yields
Theorem 1: For tracking a reference trajectory, given the the following result:
dynamic model of a WMR under partial-loss-of-effectiveness
fault and dead zone (18), the tracking errors will asymptoti- V̇ ≤ − sT Ks − ψ T Γψ. (38)
cally converge to zero as long as the proposed control scheme Applying the LaSalle invariant set theorem, we can draw
as Fig. 6 is applied where the input signals of the faulted the conclusion that s converges to zero asymptotically.
actuators and the adaptive laws are proposed as (24), (26)-
The proof finishes here.
(27), and (28), respectively.
Proof: Let us choose a candidate Lyapunov function as IV. SIMULATION
follows: In this section, one WMR is employed to simulate the
1 1 T 1 T 1 proposed FTC method. The dynamic and kinematic gains of
V = sT s + ρ̃ ρ̃ + κ̃ κ̃ + ψ T ψ. (29)
2 2α 2β 2 this WMR are shown in Table. 1.
Taking the derivative of the above equation yields the The parameters of the controller are chosen as follows:
following result: 15 0 6 0
K= , Λ= , Γ = Υ = 0.25,
1 1 0 15 0 6
V̇ =sT ṡ + ρ̃T ρ̃˙ + κ̃T κ̃˙ + ψ T ψ̇. (30) ρ = 1, ρ = 0.5, a = 10, a = 0.
α β
Page 161
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
0.8 0.8
Desired Trajectory Desired trajectory
0.6 Actual Tracjectory 0.6 Actual trajectory
0.4 0.4
Y axis (m)
Y axis (m)
0.2 0.2
0 0
-0.2 -0.2
-0.4 -0.4
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
X axis (m) X axis (m)
Fig. 7. Trajectories in the floor with the proposed FTC method. Fig. 9. Trajectories in the floor without the proposed FTC method.
0.6 0.6
e1 e1
e2 e2
0.4 0.4
0.2 0.2
0 0
-0.2 -0.2
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
Fig. 8. The evolution of the tracking errors with the proposed FTC method. Fig. 10. The evolution of the tracking errors without the proposed FTC
method.
Page 162
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.
Proceedings of 2023 International Conference on System Science and Engineering (ICSSE)
Page 163
Authorized licensed use limited to: Zhejiang University. Downloaded on November 17,2023 at 02:09:25 UTC from IEEE Xplore. Restrictions apply.