We shall address theoretical, algorithmic, and practical aspects of control for robotic systems such as manipulators and mobile rovers. Traditional control theory (mostly time-domain analysis) provides concepts, vocabulary, and mathematical models. Optimal control, adaptive control, and learning methods (e.g. neural nets) are important enhancements to be understood at a working level, but we shall minimize the traditional mathematical approach to these concepts. We shall cover recent approaches to control (computation-intensive, subsumption, fuzzy, deictic).
Simulation of complex systems is a science in itself, and is increasingly important in practical engineering. In particular, complex robotic systems are often expensive to repair, dangerous, and impractical to test in their intended environment (e.g. Mars, World War III, a nuclear reactor meltdown). Virtual reality techniques make simulation of visual stimuli possible. We expect that simulation will play a large part in the course.
Laboratory experience will consist of interaction with software simulation packages, probably writing code for real-time control (using Macintosh or Sparcstation platforms). Also possible are projects in ME or CS robotics and control laboratories. Almost certainly a significant focus of the engineering component of the course will be the twin outdoor MiniBaja vehicles to be constructed by the School of Engineering. One is to be equipped with real-time visual and control capabilities. This course can have significant impact on the design of the electro-mechanical, electronic, and software components of this vehicle.
Back to URCS Faculty directory