Abstract:

Dynamics, softness, deformability, and difficult-to-detect objects will be critical for new domains in robotic manipulation. But there are complications–including unmodelled dynamic effects, infinite-dimensional state spaces of deformable objects, and missing features from perception. This talk explores learning methods based on multi-view sensing, acoustics, physics-based regularizations, and Koopman operators and proposes a novel multi-finger soft manipulator to enable new manipulation capabilities. We demonstrate how the proposed methods can recover transparent object geometry, densely track deformable object state over time, train robot systems using vision and Koopman operators, use sound to learn a model of friction for rapid non-prehensile object transportation, manipulate hair, shape pottery, and perform high-speed non-prehensile in-hand manipulation (aka pen spinning).

Bio:

Jeff Ichnowski is an assistant professor at Carnegie Mellon University’s Robotics Institute. He was a postdoc at UC Berkeley’s Sky Computing/RISE lab, AUTOLAB, and BAIR. Before returning to academia, he was the principal architect at SuccessFactors, Inc., one of the world’s largest cloud-based software-as-a-service companies. He is a co-founder of Jacobi Robotics, Inc., a startup that builds software to power the next generation of robotics. His lab, the Momentum Robotics Lab, explores robot algorithms and systems for high-speed motion, task, and manipulation planning, using cloud-based high-performance computing, optimization, and deep learning.