Road following for autonomous vehicle navigation using a concurrent neural classifier

Abstract

The paper presents an original approach for visual identification of road direction of an autonomous vehicle using a neural network classifier called Concurrent Self-Organizing Maps (CSOM), representing a winner-takes-all collection of neural modules. We present the experimental results obtained by computer simulation of our model. The path to be identified has been quantized in 5 output directions. For training and testing the neural model, we captured and labeled a road image data set which has been divided in two lots: 30 images for training and other 30 images for test. We have also performed, trained and tested a real time neural path follower based on CSOM model, implemented on a mobile robot (car toy).