Invited talk - Multi-robot shape assembly: from theory to practice

Abstract: The fascinating collective behaviors of biological systems in nature has inspired extensive research work on multi-robot systems. Shape assembly is an important research task where a multi-robot system can form a desired geometric pattern through distributed coordination mechanisms. The shape assembly task has a unique technical challenge: the global pattern is a global constraint. This constraint creates a competitive relationship among cooperative robots. Traditional approaches tend to address these challenges through goal assignment. This talk will introduce a novel “selfless” exploration strategy for multi-robot shape assembly. Compared with the state-of-the-art methods, the efficiency of the proposed one is dozens of times higher, especially when the swarm size is large. In addition, it has good fault-tolerance abilities and can be easily extended to tasks such as complex environment exploration and collaborative transportation. This work was recently published by Nature Communications.

Bio:

Dr. Shiyu Zhao is an Associate Professor at the School of Engineering at Westlake University and the Director of the Intelligent Unmanned Systems Laboratory. He received the BE and ME degrees from Beijing University of Aeronautics and Astronautics and the Ph.D. degree in Electrical Engineering from the National University of Singapore. Before he joined Westlake University, he was a Lecturer in Aerospace Systems at the University of Sheffield, UK. Dr. Zhao’s research interests lie in the theories of estimation, control, and reinforcement learning and their applications to intelligent robotic systems especially multi-robot systems. His new textbook “Mathematical Foundations of Reinforcement Learning” will be jointly published by Tsinghua University Press and Springer Nature Press in 2024. This book has received 1.3k stars on GitHub.