SPEAKER: Dr. Changhyun Choi, Ph.D
The goal of my research is to enable robots to perceive objects reliably, to manipulate them effortlessly, and to adapt to new objects and manipulation tasks. Toward this objective, my focus is on developing new hand-eye coordination capabilities for robots, specifically 1) dependable visual perception, 2) compliant manipulation, and 3) closing the loop between them. In this talk, I will present my robot vision research which harnesses 3D shape prior and a deep learning algorithm in order to reason about manipulation relevant information. A compliant soft hand will be shown with a focus on its compliance and adaptability. I will then address closing the loop between the visual perception and the soft manipulation and present its applications to object grasping of previously unseen objects and object assembly manipulation. In the later part of my talk, I will introduce challenging yet exciting future research directions.
Dr. Choi is a Postdoctoral Associate in the Computer Science & Artificial Intelligence Lab (CSAIL) at Massachusetts Institute of Technology (MIT) working with Prof. Daniela Rus. He obtained a Ph.D. in Robotics at the School of Interactive Computing, College of Computing, Georgia Institute of Technology, wherein he was also affiliated with the Institute for Robotics and Intelligent Machines (IRIM). His research interests are in visual perception for robotic manipulation, with a focus on deep learning for object grasping and assembly manipulation, soft manipulation, object pose estimation, visual tracking, active perception, visual verification, and 3D registration.