A MOVEMENT CONTROL OF A SMALL MOBILE 3-PI ROBOT IN A MAZE USING ARTIFICIAL NEURAL NETWORK

Abstract

An intention of this article is to present a technical implementation of control of a small mobile 3-Pi robot which moving in a maze along a predefined guide line. To acquire the robotic car position information, the reflectance infra-red sensors serve as input sensors. The control of the direction of the robot's movement is performed by a single-layer neural network including two neurons. The weights (memory) of the neurons are adapting with respect to input sensors signals and the output calculated via Hebbian learning. Based on previous experiments, activation function bipolar sigmoid performs the best for a maze solving problem.