"Grand Challenges in Robotics: Dexterity" was a series of five online panels, held January - April 2023, with thought leaders exploring what may be one of the most impactful future technologies: general-purpose robotic dexterity.
Machine Learning for Robot Manipulation and Dexterity
April 19, 2023, 12-1 PM CT
Machine learning has the potential to provide the multimodal representations and adaptive skills needed to bring robot manipulation and dexterity into a wide range of new and challenging application domains. To achieve these goals, robots will need to utilize a variety of different information sources, including demonstrations, simulations, interactive perception, and trial-and-error learning. Our panel of experts will discuss the various learning challenges posed by robot manipulation, the roles of different types of learning in achieving dexterity, and future trends that will lay the foundation for ubiquitous robot manipulation.
Moderator: Oliver Kroemer, Carnegie Mellon University
Panelists: Rika Antonova, Stanford University; Chelsea Finn, Stanford University and Google Brain; Animesh Garg, University of Toronto and NVIDIA; Andy Zeng, Google
Industrial Dexterity: Rate, Range, and Reliability
April 5, 2023, 12-1 PM CT
E-commerce has increased dramatically due to the pandemic, putting tremendous pressure on warehouses to meet the demand. As each order and package varies, robots must dynamically and reliably plan grasps and motions at a high rate of speed for a broad range of packages. Warehouse applications are a "sweet spot" for emerging advances in robot sensing, planning, and control; a number of companies are using "Industrial Dexterity" to handle millions of packages per day. This panel will consider the state of the art and open problems for new research.
Moderator: Ken Goldberg, University of California, Berkeley and Ambi Robotics
Panelists: Jeff Mahler, Ambi Robotics; Matt Mason, Carnegie Mellon University and Berkshire Grey; Lauren Miller, Boston Dynamics
The Economic and Societal Impacts of the Coming Dexterity Revolution
March 2, 2023, 12-1 PM CT
What would advances in robotic dexterity mean for industry, workers, and consumers? Based on the data we have from industrial and collaborative robots, we can anticipate challenges with technology design and adoption, as well as winners and losers from any leap forward in automation. This panel will provide a multidisciplinary perspective on the impact of robots today with an eye toward how dexterity might avoid the challenges that previous generations of robots have faced.
Moderator: Julie Shah, Massachusetts Institute of Technology
Panelists: Ben Armstrong, Massachusetts Institute of Technology; Anders Beck, Universal Robots; Robert Seamans, New York University
Novel Materials for Robust Grasping and Manipulation
January 30, 2023, 3-4 PM CT
Novel material architectures capable of matching the mechanics, sensing capabilities, and articulated motions of natural human hands have the potential for transformative impact in robotic grasping and manipulation. Such advancements build on efforts to combine research in dexterous manipulation and haptics with emerging methods in soft robotics and integrated material systems. In this panel, experts from these domains will discuss and debate how progress in robotic materials is impacting the field of manipulation.
Moderator: Carmel Majidi, Carnegie Mellon University
Panelists: Elliot W. Hawkes, UC Santa Barbara; Tess Hellebrekers, Meta AI; Nancy Pollard, Carnegie Mellon University; Yon Visell, UC Santa Barbara
Robot Dexterity: The Potential and The Path Forward
January 11, 2023, 12-1 PM CT
Despite rapid progress in AI, machine learning, and certain subfields of robotics, progress in robot dexterity has been slow. Yet the potential economic and societal impact of dexterous robotic manipulation is enormous, spanning fields such as manufacturing, warehouse logistics, elder care, space exploration, and food preparation.
Moderator: Ed Colgate, Northwestern University
Panelists: Kevin, Lynch, Northwestern; Shuran Song, Columbia University; Russ Tedrake, MIT, Toyota Research Institute
Back to top