Task coordination and assistive opportunity detection via social interaction in collaborative human-robot tasks

Abstract

In most environments, task collaboration requires efficient, flexible communication between collaborators. In the case of tasks involving human-robot collaboration, the robot must effectively convey and interpret communicative actions about the current and intended state of the task environment and coordinate its behavior with those of its collaborators, human and otherwise. A framework for collaborative communication should not only allow the robot to reason about its actions in relation to those of others but should also support reasoning about the robot's own intentions and those ascribed to it by the collaborators. We present such a framework, inspired by Theory of Mind and making use of perspective taking, and show how it could be used to support several collaborative functions, including detection of opportunities to assist.