Title :
Testing the rank of the Hankel covariance matrix: a statistical approach
Author :
Camba-Mendez, Gonzalo ; Kapetanios, George
Author_Institution :
Nat. Inst. of Econ. & Social Res., London, UK
fDate :
2/1/2001 12:00:00 AM
Abstract :
The rank of the Hankel matrix, corresponding to a system transfer function, is equal to the order of its minimal state space realization. The computation of the rank of the Hankel matrix is complicated by the fact that its block elements are rarely given exactly but are estimated instead. We propose statistical tests to determine the rank of the Hankel matrix. We also provide a Monte Carlo study on the reliability of these tests compared to existing procedures
Keywords :
Hankel matrices; Monte Carlo methods; covariance matrices; discrete systems; multivariable control systems; reduced order systems; state-space methods; statistical analysis; transfer functions; Hankel covariance matrix; Monte Carlo study; block elements; minimal state space realization; rank testing; statistical approach; Adaptive control; Control systems; Cost function; Covariance matrix; Monte Carlo methods; Optimal control; Optimization methods; Signal design; Testing; Transfer functions;
Journal_Title :
Automatic Control, IEEE Transactions on
