Autonomy Talks - Guillaume Sartoretti: Distributed Collaboration in Robotic Multi-​Agent Systems

Описание: Autonomy Talks - 27/09/2021

Speaker: Prof. Guillaume Sartoretti, National University of Singapore

Title: Distributed Learning Based Scalable Collaboration in Robotic Multi-​Agent Systems

Abstract: My work has dealt with the curse of dimensionality in high degree-of-freedom (DOF) robot systems: as the number of agents (robots or DOFs) in the system grows, so does the combinatorial complexity of coordinating them. There are many solutions to managing this complexity growth, and my work has favored distributed, and more recently decentralized approaches, whether it be for a team of mobile robots or a single articulated robot. My work has embraced advances in distributed reinforcement learning to let multiple agents learn a common decentralized policy in a time-efficient manner. This has produced collaborative policies that naturally scale to an arbitrary numbers of agents, while remaining near-optimal. In this talk, I will present approaches to the problem of 1) one-shot and lifelong multi-agent path finding (e.g., for warehouse automation), 2) collective robotic construction of 3D structures, and 3) controlling the posture of an articulated robots during locomotion over steep or unstructured terrains. I will present experiments on autonomous ground vehicles and on a hexapod robot to validate some of the approaches described in this talk, and finally briefly go over some of my lab's ongoing projects.

If you want to know more, please visit https://idsc.ethz.ch/research-frazzol...