marina@ai.toronto.edu (Marina Haloulos) (10/24/89)
ACTIVITIES FOR THE WEEK COMMENCING OCTOBER 30 (SF = Sandford Fleming Building, 10 King's College Road) ------------------------------------------------------------- COLLOQUIUM SF1105, at 11:00 a.m., Tuesday 31 October 1989 Professor Rodney A. Brooks MIT Artificial Intelligence Lab Robot Beings Being a robot in a human habitat requires dealing with cluttered, unconstrained and dynamically changing environments. Most research on autonomous mobile robots assumes a static world. At best, dynamic aspects of the world are to be avoided. We report on a series of robots built over the last four years designed to operate in dynamic environments. These robots cannot be dynamically told what to do. Rather, like children and dogs, they do what is in their nature (which is determined by programs residing onboard in LEPRUMs on power up). They pursue their own activities while responding to the presence and actions of nearby people. The robots bristle with sensors. But rather than fuse the data from these sensors into a world model, the robots have many independent perceptual systems which are individually and intimately tied into behavior-generating networks of simple ocmputational elements. Each perceptual subsystem extracts only those aspects of the world which are relevant to the particular task for which it is tuned. Fusion happens closer to the motor level than the sensor level. The robots use the subsumption architecutre which is a methodology for implementing complex agents as an incrementally evolved network of augmented finite state machiens (FSMs with timers). In building such robots we have rejected the traditional AI approaches based on symbolic representations, and instead rely on large collections of carefully designed behavior generators to produce coherent complex patterns of interacting with the world.