Neural network-based fixed-time sliding mode control of pneumatic muscle-actuated robotic finger

Abstract

Pneumatic artificial muscles (PAMs) are widely used in various robotic and automation applications due to their high power-to-weight ratio and compliance. However, PAM systems are inherently highly nonlinear and subject to uncertainties, which pose difficulties for control design. This paper proposes a radial basis function neural network (RBFNN)-based fixed-time sliding mode controller (FTSMC) to address the high nonlinearity and uncertainties in PAM systems. Specifically, in order to address system lumped uncertainties, a RBFNN is employed to approximate these uncertainties, effectively improving the controller’s robustness. Furthermore, to ensure the control input remains within a certain range, unilateral constraint functions are utilized, ensuring the safety and stability of the system. The FTSMC is composed of a specially designed sliding surface that effectively enhances steady-state accuracy and achieves system stability within a fixed-time. The stability of the proposed method is rigorously substantiated through Lyapunov analysis. Finally, the simulations demonstrate that the proposed method exhibits superiority.