IJPAM: Volume 42, No. 3 (2008)
BY NUMERICAL OPTIMIZATION OF GENERALIZED
School of Electrical and Electronic Engineering
University of Nottingham
University Park, Nottingham, NG7 2RD, U.K.
Abstract.Direct minimization of the mean square error at the output of an FIR digital filter operating stationary signal and noise is considered. It is shown that the problem can be re-stated as a problem of minimizing a GRQ. The latter is solved by a numerical line search algorithm that requires calculations of only two outer matrix products per iteration. Design results are compared with those obtained by using conventional approaches. Other applications of the procedure are outlined.
Received: August 17, 2007
AMS Subject Classification: 93A25, 65F30
Key Words and Phrases: FIR digital filters, LMSE criterion, Wiener filter, overlapped signal and noise spectra, generalized Rayleigh quotient, gradient optimization
Source: International Journal of Pure and Applied Mathematics