MIT EI Seminar - Ted Adelson - Good tactile sensing lets robots do cool things

Title: Good tactile sensing lets robots do cool things.

Abstract: We are building soft sensitive robot fingers using a camera-based tactile sensing technology called GelSight. These fingers have superhuman resolution and provide accurate measurements of the contact region between the skin and the world. If you know the 3D geometry of this contact region, you can infer many important things, such as object pose, object shape, force, slip, hardness, and roughness. You can do system identification on objects, inferring their inertial and frictional properties from the way they feel in the hand. If the sensing is fast enough, you can use it in reactive control, enabling dexterous skills such as cable manipulation.

Bio: Edward ("Ted") Adelson is the John and Dorothy Professor of Vision Science at MIT, in the Department of Brain and Cognitive Sciences, and the Computer Science and Artificial Intelligence Laboratory (CSAIL). For much of his career he has worked on human vision and computer vision, on topics such as multiscale image representation, motion perception, material perception, and plenoptic imaging. In recent years he has been working on artificial tactile sensing, using an optically based touch system known as GelSight, exploring how high resolution touch can advance robotic manipulation. Professor Adelson is a member of the National Academy of Sciences, a Life Fellow of the IEEE, and the recipient of the Nakayama Medal in Vision Science.