论文信息 - Recursive solution and stability conditions for 2-D least-squares modeling and stabilization problems

Recursive solution and stability conditions for 2-D least-squares modeling and stabilization problems

Abstract Least-squares autoregressive modeling and the planar least-squares inverse (PLSI) stabilization are problems of interest in two-dimensional signal processing. We attempt to unify the treatment of these two problems by showing that both can be transformed into a one-dimensional problem and efficiently solved by means of a recursive Cholesky factorization algorithm. The development of this algorithm permits us also to introduce sufficient conditions to ensure the stability of the model or of the PLSI. Furthermore, the recursive nature of the algorithm provides insight into the choosing of the order of the model or of the PLSI, and the shape of the support for the coefficients. To illustrate the efficiency of the algorithm and the use of the stability conditions, modeling and stabilization examples are given.

Luis F. Chaparro | H. Q. Lu

[1] Rene F. Swarttouw,et al. Orthogonal polynomials , 2020, NIST Handbook of Mathematical Functions.

[2] T. Huang,et al. Stability of general two-dimensional recursive digital filters , 1978 .

[3] B.D.O. Anderson,et al. Stabilization of certain two-dimensional recursive digital filters , 1977, Proceedings of the IEEE.

[4] Luis F. Chaparro,et al. Two-Dimensional Linear Prediction Covariance Method and Its Recursive Solution , 1987, IEEE Transactions on Systems, Man, and Cybernetics.

[5] R. Schafer,et al. Two-dimensional linear prediction and its application to adaptive predictive coding of images , 1984 .

[6] Luis F. Chaparro,et al. Sufficient conditions for minimum phase two-dimensional PLSI's , 1984 .

[7] Y. Genin,et al. Two-dimensional stability and orthogonal polynomials on the hypercircle , 1977, Proceedings of the IEEE.

[8] M.N.S. Swamy,et al. A simple sufficient criterion for the stability of multidimensional digital filters , 1982, Proceedings of the IEEE.

[9] S. Treitel,et al. Stability and synthesis of two-dimensional recursive filters , 1972 .

[10] P. Karivaratha Rajan,et al. Formulation of 2-D normal equations using 2-D to 1-D form preserving transformations , 1988, IEEE Trans. Acoust. Speech Signal Process..