MLP neural networksto control a mobile robot using multiple goal points

摘要

In this paper, a new path tracking method based on an artificial neural network (ANN) is presented. The controller uses a multilayer perceptron ANN that computes the steering command to follow a previously recorded path. Instead of using a single goal point placed at a given distance of the vehicle, the proposed method considers a segment of the path ahead of the vehicle by defining several goal points. Thus, the steering command is obtained by using the set of goal points, in such a way that the characteristics of the segment of the path are taken into account. The method has been implemented in the Romeo 4R autonomous vehicle designed and built at the University of Seville. The paper presents the method and the results of the experiments with Romeo 4R