Doctor of Philosophy (PhD)
Electrical and Computer Engineering
This dissertation proposes some H-infinity and H2 performance preserving controller reduction methods for linear systems. The proposed methods can guarantee robust stability and performance for the closed-loop system with the reduced order controllers. Several H-infinity stability and performance preserving controller reduction methods are proposed in this dissertation. It is shown that the weighting functions used in the proposed controller reduction methods can be directly obtained from the parametrization of the H-infinity controllers. Hence, comparing with the most existing controller reduction approaches, the proposed controller reduction methods require less computation and are easy to apply. At the same time, several algorithms are proposed to simplify some existing controller reduction algorithms. Examples are explored to demonstrate the advantages of the proposed controller reduction methods. The parallel problems are also discussed for H2 performance preserving controller reductions. Furthermore, some parallel controller reduction methods are presented to reduce controllers for preserving the closed-loop system stability and performance. Similarly, relevant simplified algorithms are also proposed for those existing H2 performance preserving controller reduction algorithms. One example is explored to demonstrate those controller reduction methods. Another H-infinity controller reduction method is introduced for SISO system to maintain the closed-loop system stability and performance. This approach provides upper bound on the controller weighting function for general SISO H-infinity control problem, and then a lower order controller is provided using frequency weighted model reduction method, which preserves stability and performance for the closed-loop system. Finally, some possible future work are outlined.
Document Availability at the Time of Submission
Release the entire work immediately for access worldwide.
Kong, Lili, "Controller reduction for linear systems" (2012). LSU Doctoral Dissertations. 2506.