Dexterity: The Crucial Motion Intelligence for Future Robots
Abstract:
We have remarkable AIs and precise robots, but why do humans still do so much tedious physical work, from factories to households? A core missing piece is dexterity — the ability to manipulate objects skillfully. Human-level dexterity is the complex synergy of mind and body, from the lowest level of sensorimotor control to the highest level of cognitive thinking, and the crucial intelligence of task solving and execution. Achieving robot dexterity will transform industries and people’s lives, empowering robots to perform complex manual processes seamlessly in factories, households, service industries, space operations, etc.
In this talk, I will provide a deeper understanding and new perspectives on this complex problem based on my research. Why is dexterity so complex? We will broaden our understanding by studying three questions. First, what kinds of robots are made for dexterity? Second, what do complex dexterous manipulation interactions involve? Third, what skills will robots need to achieve dexterity? Finally, we will discuss the path toward general dexterity in robots.
Bio:
Xianyi Cheng is a final year Ph.D. candidate at Carnegie Mellon University, advised by Professor Matthew T. Mason. Her primary research focus is robot dexterity. Specifically, her current work focuses on the automatic generation and planning of versatile dexterous manipulation skills. She received the Foxconn Graduate Fellowship in 2018 and was selected to participate in the 2021 MIT EECS Rising Star.
Zoom Link: https://wpi.zoom.us/s/92946413217