On autonomous mobile robots and reinforcement connectionist learning
This paper discusses the possibility of constructing reactive autonomous mobile robots through reinforcement connectionist learning. The review is based on experience in solving a common task for autonomous mobile robots, that of path finding. It considers whether it is possible to use exclusively reinforcement connectionist learning techniques and what the limits of such an approach might be. It also considers the question of integration of symbolic and connectionist techniques, where traditional AI approaches seem to be necessary.
Bibliographic Reference: Paper presented: 10th European Conference on Artificial Intelligence, Workshop on Neural Networks and a New AI, Wien (AT), August 3, 1992
Availability: Available from (1) as Paper EN 36985 ORA
Record Number: 199211010 / Last updated on: 1994-12-02
Original language: en
Available languages: en