Speaker:  Yoav Sklut

M.Sc. student under the supervision of Prof. Uri Shaked.

Wednesday, May 30^th, 2018 at 15:00

Room 011, Kitot Bldg., Faculty of Engineering

A New Weighted Energy Limitation Method for H infinity Control of Linear Deterministic Systems

Abstract

In the H infinity control problem, one would like to achieve a minimum disturbance attenuation level for a given dynamic system. In many cases, the optimal H infinity state-feedback controller produces large gains that cannot be applied in practice, this mainly occurs near optimum. The common methods, currently in use, fail to address the significance of the controller's energy. In this talk, two methods are presented that address the controller energy problem. The first method provides a simple, yet effective, tool for energy limitation. The second method takes a different approach, allowing less conservative results. The two methods can be extended to weighted energy limitation, i.e. each state variable can have a different weight in the total energy of the controller. A weighted energy limitation can be applied to solve the static output-feedback problem, where a large weight is put on the direction orthogonal to the output of the system. The latter is most important in the case where the system encounters uncertainties. In this case, there is no known solution to the robust output-feedback control problem. Our solution provides not only simple static output-feedback in the uncertain case, but it also provides a means for limiting the energy of the control effort. The two methods have been developed for linear deterministic continuous systems, and are easily applied to various other problems, such as systems with state-multiplicative noise. Two numerical examples are presented that demonstrate the applicability of the theory developed.