Learning in Tele-autonomous Systems using Soar
Robo-Soar is a high-level robot arm control system implemented in Soar. Robo-Soar learns to perform simple block manipulation tasks using advice from a human. Following learning, the system is able to perform similar tasks without external guidance. Robo-Soar corrects its knowledge by accepting advice about relevance of features in its domain, using a unique integration of analytic and empirical learning techniques.
Additional InformationThis research was sponsored by grant NCC2-517 from NASA Ames and ONR grant N00014-88-K-0554. We would like to thank Karen McMahon for implementing Soar 5.0, and Mark Wiesmeyer for developing and implementing the Soar input and output interfaces. Without these extension to Soar, Robo-Soar would not have been possible. We also like to thank the staff of the Robotics Laboratory for help with the tele-autonomous software and hardware. Finally, we thank Paul Rosenbloom and Allen Newell for ideas relating to this work.