Towards an affective robot capable of being a long-term companion

Abstract

While it has been well established that affect influences judgments and decision-making, few computational models of the phenomenon exist. The work I have done and propose focuses on the role of affect in complex decision-making or judgment tasks where a pure utilitarian approach does not reflect human behavior. It is especially challenging to develop models that can predict choices made by an individual. However, as we approach having companion robots that have long-term relationships with a user, it becomes increasingly vital for the robot to be sensitive to the affect of the human and to behave in a manner that is consistent with the expectations of the user and society. The models and underlying robot architecture I describe here bring us closer to robots being accepted in our homes.