Abstract:
Robots are becoming more common in industrial manufacturing because of their speed and precision on repetitive tasks, but they lack the flexibility of human collaborators. In order to take advantage of both humans’ and robots’ abilities, we investigate how to improve the efficiency of human-robot collaborations by making sure that robots both 1. stay safe around people (i.e. avoid collisions) and 2. understand how their actions may influence people. We look at this full loop of safe human-robot interaction at multiple levels of abstraction. At a low level, we explore how a robot can use its control to improve its predictive model of a human and how this improvement may be affected by the presence of safe control. At a high level, we explore how a robot can choose its behavior such that it minimally interferes with a human collaborator’s future plans. We investigate these behaviors in both simulations and real-world human-robot interactions.

Committee:
Changliu Liu
Henny Admoni
Zachary Manchester
Benjamin Newman