Title :
A Moebius matrix representation for real symmetric Toeplitz matrices
Author_Institution :
Cirrus Logic, Broomfield, CO, USA
Abstract :
Several representations of real symmetric Toeplitz matrices are known. The Caratheodory representation is built on Vandermonde and diagonal matrices. The LU decomposition is used in linear prediction. Here a new representation of the real symmetric Toeplitz matrix is introduced as the sum of a class of Moebius matrices. The class of Moebius matrices M used here maps Toeplitz matrices T onto Toeplitz matrices via the transformation T/sub 1/=M*TM. Using the properties this class of Moebius matrices one can reformulate the problem as a Prony spectral estimation problem.
Keywords :
Toeplitz matrices; eigenvalues and eigenfunctions; matrix decomposition; parameter estimation; prediction theory; spectral analysis; Caratheodory representation; LU decomposition; Moebius matrix representation; Prony spectral estimation problem; Vandermonde matrix; diagonal matrix; eigendecomposition; linear prediction; real symmetric Toeplitz matrices; Equations; Logic; Matrix decomposition; Polynomials; Symmetric matrices;
Conference_Titel :
Signals, Systems & Computers, 1998. Conference Record of the Thirty-Second Asilomar Conference on
Conference_Location :
Pacific Grove, CA, USA
Print_ISBN :
0-7803-5148-7
DOI :
10.1109/ACSSC.1998.751555
