Human-robot collaboration within a virtual space using behavioral task morphologies

Abstract

We attempt to apply task morphologies, or the properties of commonly shared invariant task structures, between the human operator and the machine autonomies to the design of a VR-based mobile robot teleoperation system. To establish a mapping between their task structures, we analyze their continuous task performance and segment them into some qualitatively different phases using adaptive resonance theory (ART). We also discuss the effects of time-delay and discontinuity of the displayed feedback information a VR system applied to the teleoperation, and present a method for tackling this problem by deforming a VR space displayed to an operator according to his/her temporal motion properties.