A Confidence-Based Approach to Multi-Robot Learning from Demonstration

No. 1: Agents that Learn from Human Teachers

Volume

Papers from the 2009 AAAI Spring Symposium

Contents

This paper presents an overview of a series of projects exploring multi-robot learning from demonstration. We present flexMLfD, a robot independent and task independent demonstration learning system that supports a variable number of robot learners. This learning system has been fully implemented and tested, and we present three example domains, utilizing different robotic platforms, to which it has been applied. Additionally, we present scalability analysis, using up to seven real robots, examining how the number of robots being taught by the teacher at the same time affects the number of demonstrations required to learn the task, the time and attention demands on the teacher, and the delay each robot experiences in obtaining a demonstration