Learning Control of Robotic Systems

Dr. Cong Wang
Department of Electrical and Computer Engineering, NJIT


Abstract

Despite their long histories, robotics and the corresponding control techniques are once again undergoing an era of major new developments. In particular, the idea of a learning control system has become a core in a many robotic applications, ranging from manufacturing automation to robotic household service and biomedical tests and treatment. This seminar examines a variety of learning control techniques, including the relatively “old-school” learning control methods for precision motion control, to the more “avant-garde” data-oriented ideas for artificial intelligence. Topics such as iterative learning control, statistical prediction, nonparametric learning, symbolic task definition and so on will be discussed, together with their powers in various robotic applications.