- This event has passed.
Fall 2021 GRASP SFI: Matt Halm, University of Pennsylvania, “Physics-inspired learning for discontinuous contact dynamics”
October 27, 2021 @ 3:00 pm - 4:00 pm
*This was a HYBRID Event with in-person attendance in Levine 307 and Virtual attendance via Zoom
ABSTRACT
Frictional contact is the core underlying behavior of robot locomotion and manipulation, and its nearly-discontinuous dynamics make planning and control challenging even when an accurate model of the robot is available. In this talk, I will first present empirical evidence that learning an accurate model in the first place can be confounded by contact, as modern deep learning approaches are not designed to capture this non-smoothness. Second, I will discuss ContactNets, our approach which circumvents this conflict via a smooth, implicit encoding of discontinuity as signed distance functions and contact-frame Jacobians.
Presenter
Mathew Halm is a fourth-year PhD student in the Mechanical Engineering and Applied Mechanics department and the GRASP Laboratory at Penn, advised by Prof. Michael Posa. His research focuses on applying techniques from mechanics and machine learning to robotic manipulation and locomotion. Mathew is a 2019 NSF Graduate Research Fellow and a recipient of the 2020 John A. Goff Award. Prior to joining Penn, Mathew received his BS in Electrical Engineering from the University of Illinois at Urbana-Champaign.