COGNITIVE DYNAMIC NEURAL CONTROLLER DESIGN FOR MOBILE ROBOT BASED ON SELF-TUNING ON-LINE OPTIMIZATION ALGORITHM

Abstract

This paper presents the design of a cognitive dynamic neural controller (CDNC) for the trajectory tracking of non-holonomic wheeled mobile robot based on the dynamic model with self-tuning on-line optimization algorithm. The aim of the proposed controller is to solve the trajectory tracking problem of the mobile robot by finding the optimal torque control action for the two wheels of mobile robot to follow a pre-defined continuous path precisely and quickly. Particle swarm optimization (PSO) used as a fast and stable self-tuning on-line algorithm to compute the optimal parameters for the proposed controller .The robustness and effectiveness of the proposed tuning algorithm are validated by Matlab simulation results in terms of the capability of overcoming the non-representative dynamic disturbances, minimizing tracking error and obtaining the smooth and optimal torque control signals with minimum number of fitness evaluation.