Mobile Robot - Dynamic Model Controlling Using Wavelet Network

In this work, a trajectory tracking control scheme is proposed to solve trajectory tracking problems in nonlinear dynamic models of the mobile robot. The wavelet neural network (WNN) and the particle swarm optimization (PSO) algorithm are used to design the optimal motion controller for the wheeled mobile robot. The performance is measured based on the variable values of the mean square error (MSE). The work is divided into two sections; in the first section, the optimal structure is selected from wavelet neural network by changing the number of neurons (N = 10, 20, 30, 40, 50,100, 150, 200) in the hidden layer. The results are compared based on the variable values of MSE; the structure which contains the smallest value of MSE is the best choice. In the second section, the optimal structure of the fixed number of neurons, found in section one, is used for the different types of wavelet functions such as Morlet, Polywog, Shannon, and Mexican hat. The simulation of the mobile robot is specially designed to test and implement the proposed control schemes by using MATLAB Version 7.12.0.635(R2011a).