Dynamic Legged LocoManipulation: Balancing Reinforcement Learning with Model-Based Control

Sreenath, Koushil

Title:

Dynamic Legged LocoManipulation: Balancing Reinforcement Learning with Model-Based Control

dc.contributor.author	Sreenath, Koushil
dc.date.accessioned	2023-05-16T17:02:14Z
dc.date.available	2023-05-16T17:02:14Z
dc.date.issued	2023-04-12
dc.description	Presented on April 12, 2023 at 12:00 p.m. in the Marcus Nanotechnology Building, room 1116.
dc.description	Koushil Sreenath is an Assistant Professor of Mechanical Engineering, at UC Berkeley. His research interest lies at the intersection of highly dynamic robotics and applied nonlinear control. His work on dynamic legged locomotion on the bipedal robot MABEL was featured on The Discovery Channel, CNN, ESPN, FOX, and CBS.
dc.description	Runtime: 56:22 minutes
dc.description.abstract	Model-based control methods such as control Lyapunov and control barrier functions can provide formal guarantees of stability and safety for dynamic legged locomotion, given a precise model of the system. In contrast, learning-based approaches such as reinforcement learning have demonstrated remarkable robustness and adaptability to model uncertainty in achieving quadrupedal locomotion. However, reinforcement learning based policies lack formal guarantees, which is a known limitation. In this presentation, I will demonstrate that simple techniques from nonlinear control theory can be employed to establish formal stability guarantees for reinforcement learning policies. Moreover, I will illustrate the potential of reinforcement learning for more complex bipedal and humanoid robots, as well as for loco-manipulation tasks that entail both locomotion and manipulation. This brings up an intriguing question: Is reinforcement learning alone sufficient for achieving optimal results in dynamic legged locomotion, or is there still a need for model-based control methods?
dc.format.extent	56:22 minutes
dc.identifier.uri	https://hdl.handle.net/1853/71937
dc.language.iso	en
dc.relation.ispartoforgunit	Georgia Institute of Technology. Institute for Robotics and Intelligent Machines
dc.relation.ispartofseries	IRIM Seminar Series
dc.rights.metadata	https://creativecommons.org/publicdomain/zero/1.0/
dc.subject	Feedback control
dc.subject	Legged robots
dc.subject	Reinforcement learning
dc.title	Dynamic Legged LocoManipulation: Balancing Reinforcement Learning with Model-Based Control
dc.type	Moving Image
dc.type.genre	Lecture
dspace.entity.type	Publication
local.contributor.corporatename	Institute for Robotics and Intelligent Machines (IRIM)
local.relation.ispartofseries	IRIM Seminar Series
relation.isOrgUnitOfPublication	66259949-abfd-45c2-9dcc-5a6f2c013bcf
relation.isSeriesOfPublication	9bcc24f0-cb07-4df8-9acb-94b7b80c1e46