Event Date
Dr. Iman Soltani, Assistant Professor, UC Davis
Abstract
Robots still lag humans at bimanual manipulation, especially in contact-rich tasks where each hand’s action depends on the other and on what the system sees. I’ll focus on the main bottlenecks: learning policies that build coordination in as an inductive bias, controlling viewpoint to reduce occlusion (perspective control), focusing perception and compute on what matters (foveation), and pairing these with capable multi-DoF end effectors. The talk will present our latest lab results on robotic hardware for dexterous two-handed work and ML methods that integrate coordination-aware policies with active vision and foveated representations, and will close with open problems and directions for robust, scalable bimanual skills.
Bio
Iman Soltani is an assistant professor of Mechanical and Aerospace Engineering and a graduate faculty with the Departments of Computer Science and Electrical and Computer Engineering at UC Davis. Prior to joining UC Davis, he worked at the Ford Greenfield Labs in Palo Alto, CA where he led the Advanced Automation Laboratory. He received his PhD and Postdoc training at the Massachusetts Institute of Technology (MIT). His research lies at the interface of artificial intelligence, controls and machine design, focusing on the development of automation tools with applications in materials and medical research, autonomous driving and manufacturing.