New Sensors and Data-driven Methods for Robot Motor Skills
Abstract: In this talk, I will cover some recent developments and also future plans for sensor development and data-driven motor skill acquisition in Robotics. Collaboration with the Kymissis CLUE lab, we recently introduced the first tactile robot finger capable of localizing touch with sub-mm accuracy over complex multi-curved surfaces. This optics-based sensor generates a very rich signal set intended for consumption by learning algorithms. In current work, we aim to extract useful information out of this data (by finding a useful latent space) and to use it for object identification and dexterous manipulation. In a parallel project, we have recently demonstrated the use of model-free Reinforcement Learning (policy gradient methods) for joint hardware/software co-optimization of robot hands. This is an avenue of research we hope to explore much further, by jointly designing the physical hardware, sensor suite, and control policies for robots performing complex tasks.
Bio: Matei Ciocarlie is an Associate Professor of Mechanical Engineering at Columbia University. His current work focuses on robot motor control, mechanism and sensor design, planning and learning, all aiming to demonstrate complex motor skills such as dexterous manipulation. Matei completed his Ph.D. at Columbia University in New York; before joining the faculty at Columbia, he was a Research Scientist and Group Manager at Willow Garage, Inc., a privately funded Silicon Valley robotics research lab, and then a Senior Research Scientist at Google, Inc. In recognition of his work, Matei has been awarded the Early Career Award by the IEEE Robotics and Automation Society, a Young Investigator Award by the Office of Naval Research, a CAREER Award by the National Science Foundation, and a Sloan Research Fellowship by the Alfred P. Sloan Foundation.
