Learning and interacting in human-robot domains

Human-agent interaction is a growing area of research; there are many appro aches that address significantly different aspects of agent social intellig ence. In this paper, we focus on a robotic domain in which a human acts bot h as a teacher and a collaborator to a mobile robot. First, we present an a pproach that allows a robot to learn task representations from its own expe riences of interacting with a human. While most approaches to learning from demonstration have focused on acquiring policies (i.e., collections of rea ctive rules), we demonstrate a mechanism that constructs high-level task re presentations based on the robot's underlying capabilities. Second, we desc ribe a generalization of the framework to allow a robot to interact with hu mans in order to handle unexpected situations that can occur in its task ex ecution. Without using explicit communication, the robot is able to engage a human to aid it during certain parts of task execution. We demonstrate ou r concepts with a mobile robot learning various tasks from a human and, whe n needed, interacting with a human to get help performing them.