MATT MASON
Professor, Robotics Institute, and Computer Science,
Director, Robotics Institute
www

My research area is robotics. Consider a human being in its natural habitat performing its natural patterns of behavior, for instance going to the lounge for coffee. As it walks down the corridor, dodges other humans, opens doors, pours the coffee, and makes more if it took the last cup, the brain demonstrates an impressive and apparently effortless ability to (1) know what is going on out in the real world, and (2) react appropriately. That is what brains are best at; in some sense that is what brains were designed to do.

Computers cannot do it, at least not yet, not nearly as well as a human. Why not? The main problem appears to be software. We need to know how to represent information about the state of the world, how to infer such information from sensory data, and how to apply that information to choose actions. Those problems are the foundation of robotics. My research aims to explore those problems and search for the principles and techniques that will enable computers to interact with the real world.

Mobile Manipulation (A Robotic Dungbeetle). Dungbeetles use their legs both for locomotion and manipulation. We are building robots which apply the same principle. Our first system is a mobile robot which uses its wheels for manipulation as well as for locomotion. Imagine a small car planting its front wheels on a piece of paper, and using the rear wheels to drive the robot and the paper around. At the same time, if the front wheels are powered, the robot could use them to manipulate the paper. Or even more challenging: imagine the car rolling its front wheels onto a pencil, then rolling the pencil around on the desk, sort of like a dung beetle. Quicktime videos.

Dynamic Manipulation (Robot Juggling). Robots typically use static and quasistatic methods to interact with the world. People, on the other hand, are adept with dynamic methods. Some scientists have argued that the evolution of the human brain was driven by the challenges of accurate throwing. It is an interesting challenge to model-based robotics to develop robots that can exploit the dynamics of a task domain. So far we have demonstrated several instances--a snatch, a throw, and a rolling throw. Each of them is planned automatically using information about the object such as its shape and mass, and also with a good model of the dynamic behavior of our arm.