A grand engineering challenge is to develop collaborative teams of distributed, heterogeneous autonomous robots that rapidly establish situational awareness in mixed indoor/outdoor, cluttered, and unknown environments—especially when the team must primarily rely on robot-to-robot communication. Such capabilities could transform emergency response, defense, and inspection and maintenance. In this talk, I will present my lab’s physical-AI efforts to improve scalability and reliability in control and distributed coordination through: (i) morphable quadrotors that provide maneuverability, aerial manipulation, and disturbance resilience, enabling reliable operation in clutter; (ii) on-the-fly learning methods for predictive control that enable rapid adaption to unknown dynamics and disturbances with provable regret and stability; and (iii) distributed coordination algorithms that quantify the value of collaboration—who should communicate with whom, and when—to scale multi-robot coordination under communication delays and bandwidth constraints. Key in our approach is to treat the systems’ “morphology” (“body”) as an optimization variable: the robot’s structure at the single-agent level and the mesh-network topology at the multi-agent level. Then, building on bandit learning, nonlinear MPC, and submodular optimization, we co-design resource-minimal, performance-aware algorithms for adaptive coordination and control. I will present results on quadrotor hardware and in large-scale simulations (40+ robots) under realistic data-rate limitations, and conclude with open challenges.
Vasileios Tzoumas is an assistant professor at the University of Michigan, Ann Arbor (postdoc at MIT; Ph.D. at U of Pennsylvania). His research is on co-adaptive physical and artificial intelligence for scalable and reliable cyber-physical systems in resource-constrained, unstructured, and contested environments, such as robots and networked systems in defense, disaster response, and smart cities. He is a recipient of an NSF CAREER Award on networked embodied intelligence, an Army Research Office Early Career Program (ECP) award on resource-aware distributed optimization and bandit learning, the Best Paper Award in Robot Vision at the 2020 IEEE International Conference on Robotics and Automation (ICRA), an Honorable Mention from the 2020 IEEE Robotics and Automation Letters (RAL), and was a Best Student Paper Finalist Award at the 2017 IEEE Conference in Decision and Control (CDC) for a paper on robust and adaptive resource allocation and multi-agent coordination.
Amani Walker
Email