Time-Domain Wiener Adaptive Beamforming with Distributed Signal Models.

摘要

This report is a mathematical study of time-domain Wiener adaptive multichannel filtering using distributed signal models. It presents derivations for two Wiener adaptive algorithms and describes the basic technique for implementing a Wiener adaptive processor with directionally distributed signal models. In order to obtain the signal-model crosscorrelation functions needed in the Wiener adaptive algorithms, the probability density function p(tau) for the time delay between a reference sensor and the individual channels must be specified. Analytic derivations of this function are presented for inverse velocity space models, distributed ring models, and velocity-azimuth space models. These derivations are of interest in their own right: they are useful in specifying two-channel crosscorrelation functions for various directional energy distributions.