Demonstration-Based Behavior and Task Learning

Abstract

To be truly useful, robots should be able to han-dle a variety of tasks in diverse environments with-out the need for re-programming. Current sys-tems, however, are typically task-specific. Aim-ing toward autonomous robots capable of acquir-ing new behaviors and task capabilities, we de-scribe a method by which a human teacher can in-struct a robot student how to accomplish new tasks. During the course of training, the robot learns both the sequence of behaviors it should execute and, if needed, the behaviors themselves. It also stores each learning episode as a case for later general-ization and reuse.