Abstract
This paper presents a method of vision-based reinforcement learning by which a robot learns to shoot a ball into a goal. We discuss several issues in applying the reinforcement learning method to a real robot with vision sensor by which the robot can obtain information about the changes in an environment. First, we construct a state space in terms of size, position, and orientation of a ball and a goal in an image, and an action space is designed in terms of the action commands to be sent to the left and right motors of a mobile robot. This causes a “state-action deviation” problem in constructing the state and action spaces that reflect the outputs from physical sensors and actuators, respectively. To deal with this issue, an action set is constructed in a way that one action consists of a series of the same action primitive which is successively executed until the current state changes. Next, to speed up the learning time, a mechanism of Learning from Easy Missions (or LEM) is implemented. LEM reduces the learning time from exponential to almost linear order in the size of the state space. The results of computer simulations and real robot experiments are given.
Article PDF
Similar content being viewed by others
References
Bellman, R. (1957). Dynamic Programming. Princeton University Press, Princeton, NJ.
Chapman, D. & Kaelbling, L. P. (1991). “Input generalization in delayed reinforcement learning: An alogorithm and performance comparisons”. In Proc. of IJCAI-91, pages 726–731.
Connel, J. H. & Mahadevan, S. editors (1993). Robot Learning. Kluwer Academic Publishers.
Connel, J. H. & Mahadevan, S. (1993). “Rapid task learning for real robot”. In J. H. Connel and S. Mahadevan, editors, Robot Learning, chapter 5. Kluwer Academic Publishers.
Fagg, A. H., Lotspeich, D., & Bekey, G. A. (1994). “A reinforcement learning approach to reactive control policy design for autonomous robots”. In Proc. of 1994 IEEE Int. Conf. on Robotics and Automation, pages 39–44.
Inaba, M. (1993). “Remote-brained robotics: Interfacing ai with real world behaviors”. In Preprints of ISRR'93, Pitsuburg.
Kaelbling, L. P. (1993). “Learning to achieve goals”. In Proc. of IJCAI-93, pages 1094–1098.
Lin, Long-Ji (1992). Self-improving reactive agents based on reinforcement learning, planning and teaching. machine Learning, 8:293–321.
Mahadevan, S. & Connell, J. (1991) “Automatic programming of behavior-based robots using reinforcement learning”. In AAAI-'91, pages 768–773.
Mataric, M. (1994). “Reward functions for accelerated learning”. In Proc. of Conf. on Machine Learning-1994, pages 181–189, 1994.
Pomerleau, Dean A. (1993). Knowledge-based training of aritificial neural networks for autonomous robot driving. In J. H. Connel and S. Mahadevan, editors, Robot Learning, chapter 2. Kluwer Academic Publishers.
Saito, F. & Fukuda, T. (1994). “Learning architecture for real robot systems—extension of connectionist q-learning for continuous robot control domain”. In Proc. of 1994 IEEE Int. Conf. on Robotics and Automation, pages 27–32.
Sutton, R. S. (1992). “Special issue on reinforcement learning”. In R. S. Sutton(Guest), editor, Machine Learning, volume 8, pages-. Kluwer Academic Publishers.
Watkins, C. J. C. H. (1989). Learning from delayed rewards”. PhD thesis, King's College, University of Cambridge.
Whitehead, S. D. & Ballard, D. H. (1990). “Active perception and reinforcement learning”. In Proc. of Workshop on Machine Learning-1990, pages 179–188.
Whitehead, S. D. (1991). “A complexity analysis of cooperative mechanisms in reinforcement learning”. In Proc. AAAI-91, pages 607–613.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Asada, M., Noda, S., Tawaratsumida, S. et al. Purposive behavior acquisition for a real robot by vision-based reinforcement learning. Mach Learn 23, 279–303 (1996). https://doi.org/10.1007/BF00117447
Issue Date:
DOI: https://doi.org/10.1007/BF00117447