Title :
An algebraic characterization of covariance extension problem and its applications
Author_Institution :
Dept. of Syst. Sci., Osaka Univ., Japan
Abstract :
The paper considers rational covariance extension problem using impulse response sequences (Markov parameters). First the properties of the parametrization and structure of the parameter space are discussed. Second, on the parameter space we characterize important extensions and propose new interesting extensions. We show many of them (in particular maximum entropy extension and its robust variants) can be solved via semidefinite programming.
Keywords :
Markov processes; algebra; convex programming; covariance analysis; linear matrix inequalities; maximum entropy methods; parameter estimation; transient response; Markov parameters; algebraic characterization; convex programming; covariance extension problem; impulse response sequences; linear matrix inequality; maximum entropy extension; semidefinite programming; Covariance matrix; Entropy; Information filtering; Information filters; Linear matrix inequalities; MONOS devices; Random processes; Robustness; Signal processing; Speech processing;
Conference_Titel :
Decision and Control, 2003. Proceedings. 42nd IEEE Conference on
Print_ISBN :
0-7803-7924-1
DOI :
10.1109/CDC.2003.1271929