Efficient computation of tr{TR-1} for Toeplitz matrices

An efficient algorithm for the computation of tr{TR/sup -1/}, where T and R are Toeplitz matrices and R is also symmetric positive definite, is presented. The method exploits the fact that the trace of TR/sup -1/ depends only on the sum of the diagonals of R/sup -1/, and not on the whole matrix R/sup -1/. To obtain this sum, a fast efficient technique, built upon the Trench (1964) algorithm for computing the inverse of a Toeplitz matrix, is developed. The complexity of the algorithm depends on the generation function of matrix R and is O(N ln N) for generic functions and O(p ln p) for AR(p) functions.

[1]  Benjamin Friedlander,et al.  The Exact Cramer-Rao Bound for Gaussian Autoregressive Processes , 1985, IEEE Transactions on Aerospace and Electronic Systems.

[2]  P. Whittle,et al.  Estimation and information in stationary time series , 1953 .

[3]  C.-C. Jay Kuo,et al.  Spectral properties of preconditioned rational Toeplitz matrices , 1993 .

[4]  Solution of systems with Toeplitz matrices generated by rational functions , 1986 .

[5]  P. S. Ray.  A novel pulse TOA analysis technique for radar identification , 1998 .

[6]  C.-C. Jay Kuo,et al.  Design and analysis of Toeplitz preconditioners , 1992, IEEE Trans. Signal Process..

[7]  Y. Jay Guo,et al.  Channel estimation using aperiodic binary sequences , 1998, IEEE Communications Letters.

[8]  J. Lumley,et al.  Processing of Random Signals , 1978 .

[9]  W. F. Trench.  An Algorithm for the Inversion of Finite Toeplitz Matrices , 1964 .

[10]  Michael I. Miller,et al.  On the existence of positive-definite maximum-likelihood estimates of structured covariance matrices , 1988, IEEE Trans. Inf. Theory.

[11]  Benjamin Friedlander,et al.  Computation of the exact information matrix of Gaussian time series with stationary random components , 1985, 1985 24th IEEE Conference on Decision and Control.

[12]  Edward J. Wegman,et al.  Statistical Signal Processing , 1985 .

[13]  G. Strang,et al.  Toeplitz equations by conjugate gradients with circulant preconditioner , 1989 .

[14]  M.I. Miller,et al.  The role of likelihood and entropy in incomplete-data problems: Applications to estimating point-process intensities and toeplitz constrained covariances , 1987, Proceedings of the IEEE.

[15]  Toeplitz system associated with the product of a formal Laurent series and a Laurent polynomial , 1988 .