Jochem, Todd, Pomerleau, Dean
Giving robots the ability to operate in the real world has been, and continues to be, one of the most difficult tasks in AI research. Their research has been focused on using adaptive, vision-based systems to increase the driving performance of the Navlab line of on-road mobile robots. This research has led to the development of a neural network system that can learn to drive on many road types simply by watching a human teacher. This article describes the evolution of this system from a research project in machine learning to a robust driving system capable of executing tactical driving maneuvers such as lane changing and intersection navigation.
Abstract: Vision based mobile robot guidance has proven difficult for classical machine vision methods because of the diversity and real time constraints inherent in the task. But real world problems like vision based mobile robot guidance presents a different set of challenges for the connectionist paradigm. Among them are: How to develop a general representation from a limited amount of real training data, How to understand the internal representations developed by artificial neural networks, How to estimate the reliability of individual networks, How to combine multiple networks trained for different situations into a single system, How to combine connectionist perception with symbolic reasoning. Once trained, individual ALVINN networks can drive in a variety of circumstances, including single-lane paved and unpaved roads, and multi-lane lined and unlined roads, at speeds of up to 55 miles per hour.
Abstract: ALVINN (Autonomous Land Vehicle In a Neural Network) is a 3-layer back-propagation network designed for the task of road following. Currently ALVINN takes images from a camera and a laser range finder as input and produces as output the direction the vehicle should travel in order to follow the road. Successful tests on the Carnegie Mellon autonomous navigation test vehicle indicate that the network can effectively follow real roads under certain field conditions. The representation developed to perform the task differs dramatically when the network is trained under various conditions, suggesting the possibility of a novel adaptive autonomous navigation system capable of tailoring its processing to the conditions at hand.