Autonomous Teaming: Coordination, Synchronization, and Adaptation

When robots team with people, they need to be able to act in ways that are contingent, adaptive, and well-suited to the task, environment, and person. Furthermore, their actions should be understandable and useful to the context, and act in ways that afford trust. This is particularly important in safety critical settings, where they often need to do a task quickly, accurately, and safely.

We have been exploring ways we can leverage coordination and synchrony mechanisms from non-linear dynamics and physics and apply them to problems in robotics Synchrony is a fundamental aspect of nature. We see it in galaxies, ionic cell signaling in plants, economic markets, and social interaction. We can also see how these synchronous patterns emerge over time.

Our lab explores modeling this phenomena within humans teams as well as designs new algorithms to inform how robots can synthesize their behavior to cooperate with humans. Our first effort in this area was introducing a new, non-linear method to model the emergence of group entrainment (Iqbal and Riek, IEEE Trans on Affective Computing, 2015). We experimentally validated this method across both human-human and human-robot teams. We have since built anticipation algorithms for robots to sense and respond to this emergency in real time in order to coordinate their activities with people (Iqbal, Rack, and Riek, 2016, Iqbal and Riek, 2017).

We are now exploring how to extend these methods to support personalized learning and adaptation across a range of settings, particularly in home-based rehabilitation. [Woodworth, Ferrari, Zosa, and Riek, Machine Learning for Healthcare, 2018]((