The realization of a Neural Network controller for vehicle-type mobile robot navigation

Abstract

This work is part of improving the autonomous navigation of mobile robots. In an environment where many obstacles in form of wall are present, the robot must detect obstacles around it, steer to the nearest and track its deviations keeping a desired distance fixed. A joint use of information issued from the three sensors installed on the platform: in front, on the left and on the right are used to determine the proper motion for the robot at each position allowing it to navigate autonomously. The effectiveness of Neural Networks in mobile robot control is important in their learning abilities and their capacity to treat noisy data.