Block matrices with L-block-banded inverse: inversion algorithms

Authors:
A. Asif;J.M.F. Moura
Affiliations:
Dept. of Comput. Sci. & Eng., York Univ., Toronto, Ont., Canada;-
Venue:
IEEE Transactions on Signal Processing
Year:
2005

Citing 0
Cited 4

Dynamic field estimation using wireless sensor networks: tradeoffs between estimation error and communication cost

IEEE Transactions on Signal Processing
Detection of Gauss-Markov random fields with nearest-neighbor dependency

IEEE Transactions on Information Theory
A new look at state-space models for neural data

Journal of Computational Neuroscience
Efficient computation of the maximum a posteriori path and parameter estimation in integrate-and-fire and more general state-space models

Journal of Computational Neuroscience

Quantified Score

Hi-index	35.75

Visualization

Abstract

Block-banded matrices generalize banded matrices. We study the properties of positive definite full matrices P whose inverses A are L-block-banded. We show that, for such matrices, the blocks in the L-block band of P completely determine P; namely, all blocks of P outside its L-block band are computed from the blocks in the L-block band of P. We derive fast inversion algorithms for P and its inverse A that, when compared to direct inversion, are faster by two orders of magnitude of the linear dimension of the constituent blocks. We apply these inversion algorithms to successfully develop fast approximations to Kalman-Bucy filters in applications with high dimensional states where the direct inversion of the covariance matrix is computationally unfeasible.