Adaptive fault-tolerant visual control of robot manipulators using an uncalibrated camera

Abstract

In this paper, we propose an adaptive fault-tolerant visual control scheme for robotic manipulators with possible actuator failures in an uncalibrated environment. Most existing visual control approaches for robot system do not take into account actuator failures, which may prominently affect the transient performance of the system in practice. In order to moderate the detrimental and adverse effects of actuator failures on the system, a new adaptive algorithm is proposed to compensate stuck-type failures occurred in actuators. Moreover, by proposing a decoupling method, the uncertain parameter model of actuator failure is successfully separated from the dynamics model while the two models are coupled in most existing results [e.g., (Rugthum and Tao in Robotica 34(7):1529–1552, 2016; Rugthum in: International conference on engineering, applied sciences, and technology (ICEAST), pp 1–4, 2018)]. And the stability of the dynamic system and the convergence of the image error are proved by the Lyapunov analysis method. Finally, the effectiveness of the proposed control scheme is verified by comparing and analysing the tracking performance of the 3-DOF manipulator under different failure parameters.