What Lies Ahead? Expectation Management in Human-Robot Collaboration.

摘要

We aim to build robots that go beyond command-and-response and can engage in fluent collaborative behav-ior with their human counterparts. This paper discusses one aspect of collaboration fluency: expectation man-agement — predicting what a human collaborator will do next and how to act on that prediction. We pro-pose a formal time-based collaborative framework that can be used to evaluate this and other aspects of col-located human-robot teamwork, and show how expec-tation management can enable a higher level of flu-ency and improved efficiency in this framework. We also present an implementation of the proposed theoret-ical framework in a simulated human-robot collabora-tive task.