Learning from Humans for Adaptive Interaction

Abstract

Robots that will cooperate (or even compete) with humans should understand their goals and preferences. Humans leak and provide a lot of data, e.g., they take actions to achieve their goals, they make choices between multiple options, they use language or gestures to convey information. And we, as humans, are usually very good at using all these available information: we can easily understand what another person is trying to do just by watching them for a while. The goal of my research is to equip robots with the capability of using multiple modes of information sources. For this, I propose using a Bayesian learning approach, and show how it is useful in a variety of applications ranging from exoskeleton gait optimization to traffic routing.