Neural network-based recognition of mine environments

Abstract

This project is a part of a general study of exploratory navigation by a vision-guided mobile robot manoeuvring in a large, unknown, dynamic environment such as an underground mine complex. Since the robot navigation is based on intersections, our problem is make the robot to learn and to recognize images representing intersections based on a gray-scaled 360/spl deg/ panoramic view. We propose a multilayer neural network to make associations between these representations and indexes corresponding to the encountered intersections.