Human-robot collaboration for a shared mission

Abstract

We are interested in collaboration domains between a robot and a human partner, the partners share a common mission without an explicit communication about their plans. The decision process of the robot agent should consider the presence of its human partner. Also, the robot planning should be flexible to human comfortability and all possible changes in the shared environment. To solve the problem of human-robot collaboration with no communication, we present a model that gives the robot the ability to build a belief over human intentions in order to predict his goals, this model counts mainly on observing the human actions. We integrate this prediction into a Partially Observable Markov Decision Process (POMDP) model to achieve the most appropriate and flexible decisions for the robot.