Neural Active Disturbance Rejection Adaptive Lateral Manipulation Control Method for Unmanned Driving Robot

Abstract

In this article, a neural active disturbance rejection adaptive lateral manipulation control method for an unmanned driving robot (UDR) is proposed to realize accurate and stable steering and path tracking. Combined with a model of the manipulated vehicle and steering manipulator, an integrated dynamics model of the vehicle manipulated by a UDR is established. Taking the error of the body heading angle and the lateral error of the vehicle as input, an active disturbance rejection controller is designed; it includes a tracking differentiator, nonlinear state error feedback (NLSEF) device, and an extended state observer. To achieve a better performance, the combination mode of the NLSEF is adjusted adaptively by a radial basis function NN. The network is then initialized by a particle swarm optimization algorithm. Finally, the results of simulations and experiments show that the proposed method effectively improves the performance of stable steering and path tracking of the UDR.