Overview
Explore a comprehensive lecture on optimizing robot actions in human environments, delivered by Anca Dragan from UC Berkeley's Electrical Engineering and Computer Sciences department. Delve into the complexities of human-robot interaction, covering topics such as human prediction, autonomous driving, and shared utility. Learn about the challenges of predicting human behavior, the limitations of the "predict and react" approach, and innovative strategies for coupled planning. Examine real-world applications in autonomous driving, including navigating four-way stops and incentivizing human cooperation. Gain insights into the nuances of human-human interaction and how these principles can be applied to improve robot-human communication and collaboration.
Syllabus
Intro
Three Kinds of People
Human Prediction Problem
Autonomous Driving Example
Predict and React Approach
Human Human Interaction
Optimal trajectory
Fourway stops
Moving backwards
Incentivizing people
Communication
Coupled Planning
People Dont Always Plan
Shared Utility
Taught by
Simons Institute