|
| |
| M.D. Schmill, M.T. Rosenstein, P.R. Cohen, and P.E. Utgoff. Learning what is relevant to the effects of actions for a mobile robotIn Proceedings of the Second International Conference on Autonomous Agents, 247-253, 1998. | |
| Abstract: We have developed a learning mechanism that allows robots to learn the conditional effects of their actions. Based on sensiomotor experience, a robot can explore it's environment, classifying the observed effects of its actions on its sensor readings. These observations can then be used to form a context operator difference table, a structure that relates contexts and actions to their effects on the environment. This table can then be the basis for learning more complex behaviors. We have applied this learning mechanism to the Pioneer 1 mobile robot with promising results. | |
| Download: pdf (286KB), ps.gz (164KB) |
|
updated 05-Sep-2000 mtr@cs.umass.edu |