Passivity-Based Control and Optimization in Networked Robotics -From Human-Swarm Collaborations to Distributed Optimization-

March 04, 2016, Webb 1100

Takeshi Hatanaka

Tokyo Institute of Technology, Controls Center of Engineering

Abstract

In this talk, we highlight the role of passivity in human-swarm collaborations and distributed optimization. The former half addresses human operator enabled motion synchronization of an ensemble of kinematic robots to desired reference motion under distributed information exchanges among the robots and between the operator and robots. A feedback loop connecting the operator and the robots is designed based on passivity, and asymptotic synchronization is then demonstrated by assuming passivity of a human operator decision process. The aforementioned passivity assumption is further studied through experiments on a human-in-the-loop simulator and closed-loop system identification. In the later half, it is revealed that the control architecture presented above also provides a solution to a class of the distributed optimization problems by just replacing the human operator block by a collection of the gradients of local cost functions. The passivity paradigm allows us not only to prove convergence to the optimal solution but also to integrate the other passive blocks while ensuring the convergence property. Finally, we show an application of the presented solution to 3-D human localization problem using pedestrian detection techniques.

Speaker's Bio

Takeshi Hatanaka received the Ph.D. degree in applied mathematics and physics from Kyoto University, in 2007. Since 2007, he has been with Tokyo Institute of Technology, where he is currently an associate professor. His research interests include networked robotics and energy management systems. He is the coauthor of ”Passivity Based Control and Estimation in Networked Robotics” (Springer, 2015). He received 2014 Outstanding Research Award and 2009, 2015 Outstanding Paper Award from SICE, and 10th Asian Control Conference Best Paper Prize Award. He is a member of the Conference Editorial Board of IEEE CSS.