Performance Analysis of Adaptive Algorithms for Envelope-Constrained Filtering

Abstract

The performance of the envelope-constrained filter designed by using the newly developed adaptive algorithms will deteriorate when the input signal is contaminated by noise. Two average window techniques are described and then applied to reduce the effect of noise. It is demonstrated that due to use of an averaging scheme, the convergence of the adaptive algorithms can be greatly improved, the misadjustment caused by the noise can be reduced considerably, and the envelope constraints are more easily satisfied by the filter weight estimate, Simulation results are given.