Adaptive variable-gain sliding mode control of robot manipulators with full state constraints

This paper introduces an adaptive variable gain sliding mode control strategy tailored for robotic manipulators with uncertainties and system constraints. A novel sliding surface is initially crafted to handle state constraints, where the gain is adjusted based on the proximity of system states to predefined constraints. Subsequently, an adaptive sliding mode control (SMC) scheme is formulated to ensure tracking error convergence. A neural network (NN) is harnessed to estimate uncertainties, with a composite learning approach developed to expedite NN weight updates. The Lyapunov framework validates the closed-loop system’s stability, and simulations underscore the efficacy of the proposed identification and control algorithms.