Enhance Control Performance of a Pneumatic Artificial Muscle System Using RBF-Neural Network Approximation and Power Rate Exponential Reaching Law Sliding Mode Control

Abstract

This research focuses on the integration of a Radial-Basis-Function Neural Network (RBFNN) for uncertainty approximation in Pneumatic Artificial Muscle (PAM) systems within the framework of Power Rate Exponential Reaching Law Sliding Mode Control (PRERL-SMC). Configured in an antagonistic manner, PAMs provide a range of benefits for developing actuators with human-like characteristics. Nevertheless, their intrinsic nonlinearity and uncertain behavior are obstacles to attaining accurate control, particularly in rehabilitation scenarios where ensuring control precision is imperative for safety and effectiveness. The proposed method leverages a power rate exponential reaching law to ensure chattering-free control and swift convergence towards desired trajectories, while the RBFNN effectively approximates system uncertainties. Through comprehensive experiments, we compare the RBF-PRERL-SMC approach with conventional control methods, showcasing its superior performance in tracking various trajectories. Notably, our strategy proves robust against external perturbations, demonstrating its applicability in rehabilitation scenarios.