Adaptive Sliding Mode Control of Robot Manipulators with System Failures

Rui Li,Yong Chen,Guanyu Lai,Liang Yang

doi:10.3390/math10030339

Abstract

This paper presents a novel adaptive sliding mode controller for a class of robot manipulators with unknown disturbances and system failures, which can well achieve the asymptotic tracking, and avoid some possible singularity problems. A new virtual controller is designed such that the chosen Lyapunov function can be transformed into a non-Lipschitz function, based on which, the system states can arrive at the specified sliding surface within a finite time regardless of the existence of system failures/faults. By fusing an integral fast terminal nonsingular SMC and a robust adaptive technique, the tracking error can be steered into a preset range in a set time and some possible singularity problems are avoided elegantly. With our proposed scheme, the loss coefficient is well estimated, and the stability of the system can be guaranteed even in the presence of the total loss of actuator outputs. The experiment and simulation results are presented to illustrate the effectiveness of the proposed control scheme.

Highlights

Nowadays, fault-tolerant control (FTC) has drawn a great deal of attention from many researchers
Actuator failure can be roughly categorized into two classes: (1) partial loss of effectiveness (PLOE), and (2) total loss of effectiveness (TLOE)
Inspired by the prior works mentioned above, we propose a novel integral nonsingular fast terminal sliding mode control (INFTSMC) for a robot with unknown disturbance and system faults, and the proposed approach can provide a finite-time convergence without singular problem, fast transient response, and high tracking precision

Summary

Introduction

Fault-tolerant control (FTC) has drawn a great deal of attention from many researchers. Inspired by the prior works mentioned above, we propose a novel integral nonsingular fast terminal sliding mode control (INFTSMC) for a robot with unknown disturbance and system faults, and the proposed approach can provide a finite-time convergence without singular problem, fast transient response, and high tracking precision. (1) A new virtual control scheme is constructed, based on which the continuous Lyapunov function can be transferred into the non-Lipschitz function With this non-Lipschitz function, the system states can arrive at the specified sliding surface within a finite time even in the presence of actuator failure.

Choose a Lyapunov candidate function as

Adaptive parameter

Conclusions