Mood as an affective component for robotic behavior with continuous adaptation via Learning Momentum

Abstract

The design and implementation of mood as an affective component for robotic behavior is described in the context of the TAME framework - a comprehensive, time-varying affective model for robotic behavior that encompasses personality traits, attitudes, moods, and emotions. Furthermore, a method for continuously adapting TAME's Mood component (and thereby the overall affective system) to individual preference is explored by applying Learning Momentum, which is a parametric adjustment learning algorithm that has been successfully applied in the past to improve navigation performance in real-time, reactive robotic systems.