Adaptive Predefined-Time SMC for Uncertain Robotic Systems

Abstract

This paper presents a novel adaptive predefined-time sliding-mode control (SMC) strategy for the tracking problem of uncertain robotic manipulators. Firstly, the unknown dynamic parameters and the bound vector of the disturbances are integrated into a compounded uncertainty. Considering that the boundary of the compounded uncertainty is not readily available in practical applications, the adaptive neural networks (NNs) are adopted to compensate for the compounded uncertainty of the robot. Moreover, to enhance the robustness and accelerate the system convergence, a novel nonsingular sliding surface approach is integrated into the controller design. Then, an NNs-based predefined-time controller is proposed to guarantee that the actual trajectory tracks the reference trajectory, allowing the maximum convergence time to be pre-adjusted using explicit parameters without losing robustness property. At last, simulation results conducted with a two-joint manipulator confirm both the exceptional performance and effectiveness of the developed method.