Toward Intelligence for Humanoid Laborers: From Planning to Whole-Body Control

Abstract:

Human physical labor has long been a foundation of civilization, yet our ability to supply it has always been limited.  Humanoid robots have the potential to dramatically expand the supply of human-like physical labor, but realizing this vision will require substantial advances in robot intelligence. In this talk, I will discuss our research toward developing the intelligence needed for general-purpose humanoid labor, involving both individual robots and teams of robots working together. I will organize the discussion around three interconnected components of a humanoid robotics software stack: whole-body control, learned skills, and high-level planning. Through examples of our recent work, I will highlight progress, open challenges, and questions about how these components should interact to enable general-purpose humanoid labor.

Speaker Bio: 

Alan Fern is a Professor of Computer Science in the School of EECS at Oregon State University (OSU) and a Fellow of the Association for the Advancement of Artificial Intelligence (AAAI). He received his Ph.D. in Computer Engineering from Purdue University in 2004 and his B.S. in Electrical Engineering from the University of Maine in 1997. His research spans a variety of topics in artificial intelligence and robotics, with a particular emphasis on building systems that learn from experience. He directs the Dynamic Robotics and AI Laboratory (DRAIL), which studies AI approaches for enabling robots to perform whole-body work in the real world. He is also the Director of AI for the AgAID National AI Institute.