MANIAC: A Next Generation Neurally Based Autonomous Road Follower - Robotics Institute Carnegie Mellon University

MANIAC: A Next Generation Neurally Based Autonomous Road Follower

Todd Jochem, Dean Pomerleau, and Chuck Thorpe
Conference Paper, Proceedings of 3rd International Conference on Intelligent Autonomous Systems (IAS '93), February, 1993

Abstract

The use of artificial neural networks in the domain of autonomous vehicle navigation has produced promising results. ALVINN [Pomerleau, 1991] has shown that a neural system can drive a vehicle reliably and safely on many different types of roads, ranging from paved paths to interstate highways. Even with these impressive results, several areas within the neural paradigm for autonomous road following still need to be addressed. These include transparent navigation between roads of different type, simultaneous use of different sensors, and generalization to road types which the neural system has never seen. The system presented here addresses these issue with a modular neural architecture which uses pretrained ALVINN networks and a connectionist superstructure to robustly drive on many different types of roads.

Notes
Also appears in the Proceedings of the Image Understanding Workshop, April 1993, Washington D.C., USA

BibTeX

@conference{Jochem-1993-13449,
author = {Todd Jochem and Dean Pomerleau and Chuck Thorpe},
title = {MANIAC: A Next Generation Neurally Based Autonomous Road Follower},
booktitle = {Proceedings of 3rd International Conference on Intelligent Autonomous Systems (IAS '93)},
year = {1993},
month = {February},
}