Title :
Employing Kronecker Canonical Form for LMI-Based
Synthesis Problems
Author_Institution :
Dept. of Electr. Eng., Linkoping Univ., Linkoping, Sweden
Abstract :
The Kronecker canonical form (KCF) can be employed when solving H∞ synthesis problem. The KCF structure reveals variables that can be eliminated in the semidefinite program that defines the controller. The structure can also be used to remove states in the controller without sacrificing performance. In order to find the KCF structure, we can transform the relevant matrices to a generalized upper triangular (Guptri) form using orthogonal transformations. Thus, we can avoid finding the KCF structure explicitly, which is a badly conditioned problem.
Keywords :
H∞ control; control system synthesis; linear matrix inequalities; mathematical programming; matrix algebra; Guptri; KCF structure; Kronecker canonical form; LMI-based H∞ synthesis problems; generalized upper triangular form; matrices; orthogonal transformations; semidefinite program; Benchmark testing; Closed loop systems; Linear matrix inequalities; Null space; Riccati equations; Sensitivity; Singular value decomposition; Controller synthesis; Kronecker canonical form (KCF); linear matrix inequality (LMI); model reduction; semidefinite programming; singular value decomposition (SVD);
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.2011.2179846