The patent badge is an abbreviated version of the USPTO patent document. The patent badge does contain a link to the full patent document.
The patent badge is an abbreviated version of the USPTO patent document. The patent badge covers the following: Patent number, Date patent was issued, Date patent was filed, Title of the patent, Applicant, Inventor, Assignee, Attorney firm, Primary examiner, Assistant examiner, CPCs, and Abstract. The patent badge does contain a link to the full patent document (in Adobe Acrobat format, aka pdf). To download or print any patent click here.
Patent No.:
Date of Patent:
Feb. 10, 1998
Filed:
Oct. 20, 1995
Mark S Spurbeck, Louisville, CO (US);
Richard T Behrens, Louisville, CO (US);
Cirrus Logic, Inc., Fremont, CA (US);
Abstract
A method and apparatus for computing, in real time, the coefficients C.sub..tau. (n) of a time varying FIR filter according to an optimum singular value decomposition (SVD) algorithm. The coefficients of a plurality of FIR filters are represented by a M.times.N matrix A.sub.M.times.N, where M is the number of FIR filters and N is the number of coefficients in the impulse response of each FIR filter (i.e., the number of filter taps). The A.sub.M.times.N matrix is factored into F.sub.M.times.N and G.sub.N.times.N matrices, and a singular value decomposition of the A.sub.M.times.N matrix is computed as A.sub.M.times.N =U.sub.M.times.N .cndot.D.sub.N.times.N .cndot.V.sub.N.times.N, where U.sub.M.times.N is a M.times.N unitary matrix, D.sub.N.times.N is a N.times.N diagonal matrix {.sigma..sub.1, .sigma..sub.2, . . . , .sigma..sub.N }, .sigma..sub.i are the singular values of A.sub.M.times.N (and .sigma..sub.1 .gtoreq..sigma..sub.2 . . . .gtoreq..sigma..sub.N .gtoreq.0), and V.sub.N.times.N is a N.times.N unitary matrix. By selecting a predetermined L number of the largest singular values .sigma. to generate a reduced size diagonal matrix D.sub.L.times.L, it is possible to extract the first L columns from the U.sub.M.times.N matrix to form a reduced U.sub.M.times.L matrix and to extract the first L rows from the V.sub.N.times.N matrix to form a reduced V.sub.L.times.N matrix. In this manner the filter matrix A.sub.M.times.N can be approximated by reducing the F.sub.M.times.N and G.sub.N.times.N matrices by replacing N with L such that: