Task-Relevant Perception for Robot Grasping and Manipulation (WPI Only)
Friday, 10/3/2014 2:00 PM-3:00 PM
Higgins Laboratories, 102
CCIS, Northeastern University
Abstract: Robust robotic grasping and manipulation is an important enabling capability needed to develop robots that can perform useful work in the real world. However, robust grasping often depends upon the ability of the robot to accurately perceive relevant state of objects in the world around it. In this talk, I will review recent work in my research group that addresses this challenge. In particular, I will talk about an approach to grasping novel objects that is based on
localizing grasp affordances in the environment. This approach is interesting because it can grasp robustly, even when the environment is unstructured and the robot has no prior knowledge about the geometry of the objects to be grasped. I will also describe a new approach to using tactile sensing to do small parts manipulation and insertion. Our method applies computer vision algorithms in the tactile sensing domain. All of this work has been demonstrated and evaluated on real world robot platforms.
Bio: Robert Platt is an assistant professor in the College of Computer and Information Science at Northeastern University. Prior to that, he was an assistant professor at SUNY Buffalo, a research scientist at MIT, and a technical lead on the Robonaut project at NASA JSC. He holds a Ph.D. from the University of Massachusetts, Amherst.
For more information, e-mail email@example.com.