On the study of human-robot collaboration

Abstract

Human-robot interaction (HRI) is being explored with the aim of creating rich, effective and satisfying interactions between humans and robots. Human-robot collaboration for learning, in which the measurable outcomes are information gain on the part of the humans in the system, is specifically focused upon. In the context of this focus on collaboration for learning, rich means a cognitively sophisticated interaction in which humans and robots communicate as peers; effective means that formal measures of human learning should yield significant outcomes; satisfying means that humans should find the interaction both useful and pleasurable. Two themes have dominated HRI research projects: how to field interactive robot systems to evaluate human-robot collaboration, and how to explicitly model human cognition and learning resulting from robot interaction. Models of interaction and evaluation methodology from human factors, HCI and cognitive psychology, have been joined and extended. Research in reasoning and social policy addresses the issue of decision-making and execution in the context of rich interaction and partial information. Research results and the evaluation methodology are validated, demonstrating significant improvement in interaction effectiveness from contributions as diverse as robot morphology design, diagnostic transparency, and affective behavior modeling. The current research thrust is a special case of collaboration: educational robotics, where robots collaborate with humans to promote learning. Statistical significance across learning themes is demonstrated using formal coding analyses. The present ambition in educational robotics is the principled design of a robotic platform and curriculum, based on an extension of coded learning themes, that can have significant impact on technology literacy and interest in science and engineering education at the middle school and high school levels