An Improved Sliding Mode Control Algorithm for Integrated Driving and Sensing Electromagnetic Actuators

Abstract

The design of control algorithms in a multi-coupling physical field is critical for advancing the application of integrated driving-and-sensing electromagnetic actuators. Accordingly, we propose a discrete fast terminal sliding mode control (FTSM) algorithm enhanced with a radial basis function neural network and delayed disturbance compensation (D-RBF-FTSM). Specifically, a discrete fast terminal sliding surface is first designed to ensure nonlinear system performance. Next, a radial basis function neural network is employed to approximate model uncertainties, with composite weight updates based on position states and the sliding surface to suppress chattering. Finally, a disturbance compensation scheme is introduced to address neural network approximation errors and external disturbances, thereby reducing boundary-layer errors and enabling high-precision control. Experimental results show that, compared with conventional discrete FTSM, D-RBF-FTSM reduces the average error by 15.2%–50.9% and the rolling standard deviation by 13.4%–49.3% across multiple loaded sinusoidal tests, demonstrating enhanced control accuracy and improved dynamic performance.