Neural network control system for a tracked robot

Abstract

In this paper the designing of a tracked robot's neural network control system is considered. The control system embodies a black line following algorithm, which is using two infrared reflector sensors for black line recognition. The neural network regulator is designed in Matlab/Simulink using the Real-Time Windows Target Toolbox. With the purpose of the neural network regulator training, course passage results of the robot with a fuzzy regulator are used.