Algorithm research and real-time simulation of neural network sliding mode position control

Abstract

This paper presents a neural network sliding mode control algorithm for position control of modular robot. This method adopts BP neural network to approximate the functional relation between the sliding hyperplane and the exponential approximation rate. At the same time, the saturation function of sliding mode control algorithm is replaced by a hyperbolic tangent function to realize the boundary design method of the sliding mode control. The results of real-time simulation show that the algorithm proposed in this paper has the merits of fast response, strong robustness, and reducing the chattering of sliding mode control. This method solves the problems that conventional PID algorithm can't solve under some circumstances, such as complicated environment, great load change, etc.