SIGNAL PROCESSING APPARATUS USING ALGORITHM FOR MINIMIZING MEAN SQUARE ERROR

摘要

PROBLEM TO BE SOLVED: To realize stable converging operation in a signal processing appara tus, using an algorithm minimizing mean square errors.;SOLUTION: Multipliers 1-1 to 1-M multiply input signals X_1(n)-X_m(n) by adaptive weights W_1(n)-W_m(n). Adder 2 combines the outputs from the multipliers 1-1 to 1-M. An error function deriving section 3 computes an error function e(n), expressing the difference between the combined value Y(n) and a reference signal r(n). An adaptive weight updating section 4 then computes adaptive weights W_1(n)-W_m(n), based on the error function e(n). The error function deriving section 3 comprises a correction coefficient generating section 11 for generating correction coefficients α, β based on the adaptive weights W_1(n)-W_m(n), a calculation section 12 for multiplying the combined value Y(n) by the correction coefficient β, a calculation section 13 for multiplying the reference signal r(n) by the correction coefficient α, and a subtracter 14 for computing the difference between the outputs of the operation sections 12 and 13.;COPYRIGHT: (C)2003,JPO