Explore dexterous manipulation and robust control in this 39-minute conference talk by MIT/Toyota Research Institute's Russ Tedrake at the Workshop on New Directions in Reinforcement Learning and Control. Delve into category-level manipulation, policy representation, and evaluation techniques for open-world manipulation. Compare robust control approaches with industry standards, examining falsification and simulation methods. Gain insights into future directions for advancing perception-based control systems in robotics and automation.
Robust Control with Perception in the Loop- Towards Open-World Manipulation - Russ Tedrake
Institute for Advanced Study via YouTube
Overview
Syllabus
Introduction
Dexterous Manipulation
Category Level Manipulation
Policy Representation
Evaluation
Robust Control vs Industry
falsification
simulation
future directions
Taught by
Institute for Advanced Study