Shortening of paraunitary matrices obtained by polynomial eigenvalue decomposition algorithms

Corr, Jamie and Thompson, Keith and Weiss, Stephan and Proudler, Ian K. and McWhirter, John G. (2015) Shortening of paraunitary matrices obtained by polynomial eigenvalue decomposition algorithms. In: 5th Conference of the Sensor Signal Processing for Defence, 2015-07-09 - 2015-07-10, Royal College of Physicians of Edinburgh. (https://doi.org/10.1109/SSPD.2015.7288523)

[thumbnail of Corr-etal-SSPD-2015-Shortening-of-paraunitary-matrices-obtained-by-polynomial]
Preview
Text. Filename: Corr_etal_SSPD_2015_Shortening_of_paraunitary_matrices_obtained_by_polynomial.pdf
Accepted Author Manuscript

Download (253kB)| Preview

Abstract

This paper extends the analysis of the recently introduced row-shift corrected truncation method for paraunitary matrices to those produced by the state-of-the-art sequential matrix diagonalisation (SMD) family of polynomial eigenvalue decomposition (PEVD) algorithms. The row-shift corrected truncation method utilises the ambiguity in the paraunitary matrices to reduce their order. The results presented in this paper compare the effect a simple change in PEVD method can have on the performance of the paraunitary truncation. In the case of the SMD algorithm the benefits of the new approach are reduced compared to what has been seen before however there is still a reduction in both reconstruction error and paraunitary matrix order.