Share Email Print
cover

Proceedings Paper

A Novel Vlsi System Of Linear Equations Solver For Real-Time Signal Processing
Author(s): Kishan Jainandunsing; Ed F.A. Deprettere
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In this paper we present a novel approach towards the problem of solving sets of linear equations, as they appear in many digital signal processing problems. This approach avoids a back substitution step or an orthogonal transformation, after the factorization step, as is the case for the conventional direct methods of QR and LQ factorization. In fact, the novel algorithm enables one to calculate the solution x forwardly from the factorization of the matrix A, using orthogonal or J-orthogonal transformations. It can be combined with the (generalized) Schur algorithm, which does the Cholesky factorization of the matrix A efficiently in case A is positive definite, symmetric and Toeplitz or close to Toeplitz. In case A is a general (non singular) matrix, the complete equations solver appears to be an orthogonal equivalent of Faddeeva's and can be similarly generalized towards a larger class of matrix arithmetic.

Paper Details

Date Published: 23 March 1986
PDF: 8 pages
Proc. SPIE 0698, Real-Time Signal Processing IX, (23 March 1986); doi: 10.1117/12.976240
Show Author Affiliations
Kishan Jainandunsing, Delft University of Technology (The Netherlands)
Ed F.A. Deprettere, Delft University of Technology (The Netherlands)


Published in SPIE Proceedings Vol. 0698:
Real-Time Signal Processing IX
William J. Miceli, Editor(s)

© SPIE. Terms of Use
Back to Top