Neural Controller for Mobile Multipurpose Caterpillar Robot

Abstract

This paper presents an advanced intelligent approach to solve the control problem of a caterpillar mobile robot (MR) able to move on inclined and vertical ferromagnetic surfaces. The authors propose structure and synthesis method of the neural controller of the MR's spatial motion for chosen robot's construction and in accordance with it's mathematical model. The peculiarity of the proposed structure lies in combination of separate speed and angle controllers in a single complex neural controller in interdependent two-channels structure of MR's control system. The synthesis method of the proposed controller is developed on the basis of genetic algorithm using complex fitness function considering quality of control for both output coordinates (speed and angle). Simulation results confirm the high efficiency of the proposed approach.