We describe an effort to train a RoboCup soccer-playing agent playing in the Simulation League by capturing data from existing players, and using it in a real-time scene recognition system. When observing a simple rule-based, stateless agent, the trained player appears to imitate the behaviour of the original. Apart from some parameter selections, the process requires little human intervention. Copyright

Additional Metadata
Conference 2006 AAAI Workshop
Citation
Lam, K. (Kevin), Esfandiari, B, & Tudino, D. (David). (2006). A scene-based imitation framework for RoboCup clients. Presented at the 2006 AAAI Workshop.