Towards autonomous robot control via self-adapting recurrent networks
This paper introduces a connectionist architecture for autonomous robot control in which second-order recurrent connections are used to provide a flexible, context-dependent mapping from percepts to actions in order to allow the network to adapt its behaviour continually to its current context and internal state. It is argued that this mechanism, to a higher degree than modular approaches, allows the robot to acquire and adapt complex behaviour autonomously.
Unable to display preview. Download preview PDF.
- Ackley, D. H. & Littman, M. L. (1990) Generalization and scaling in reinforcement learning, in Touretzky, D. S. (ed.) Advances in Neural Information Processing Systems, 550–557, Morgan Kaufmann, San Mateo, CAGoogle Scholar
- Beer, Randall D. (1995) A dynamical systems perspective on agent-environment interaction, Artificial Intelligence, 72, 173–215Google Scholar
- Brooks, Rodney (1986) A robust layered control system for a mobile robot, IEEE Journal of Robotics and Automation, 2(1), 1986, 14–23Google Scholar
- Elman, J. (1990) Finding Structure in Time, Cognitive Science, 14, 179–192Google Scholar
- Jordan, Michael I., Jacobs, Robert A. (1995) Modular and hierarchical learning systems, in Arbib, M. (ed.) The Handbook of Brain Theory and Neural Networks, MIT Press, Cambridge, MAGoogle Scholar
- Meeden, Lisa (1996) An Incremental Approach to Developing Intelligent Neural Network Controllers for Robots, to appear in IEEE Transactions on Systems, Man, and Cybernetics, 26, special issue on Learning Autonomous RobotsGoogle Scholar
- Michel, O. (1995) Khepera Simulator version 1.0 — User Manual, University of Nice-Sophia Antipolis, Valbonne, France, http://wwwi3s.unice.fr/∼om/khep-sim.htmlGoogle Scholar
- Mondada, F., Franzi, E., & Ienne, P. (1993) Mobile robot miniaturisation: A tool for investigation in control algorithms, in Third International Symposium on Experimental Robotics, Kyoto, JapanGoogle Scholar
- Pollack, Jordan B. (1991) The induction of dynamical recognizers, Machine Learning, 7, 227–252Google Scholar
- Schyns, Phillippe G. (1991) A Modular Neural Network Model of Concept Acquisition, Cognitive Science, 15, 461–508Google Scholar
- Steels, Luc (1995) When are robots intelligent autonomous agents?, Robotics and Autonomous Systems Google Scholar
- Ziemke, Tom (1996) Towards Adaptive Behaviour System Integration using Connectionist Infinite State Automata, to appear in Proceedings of the Fourth International Conference on the Simulation of Adaptive Behavior, MIT Press, Cambridge, MAGoogle Scholar