Using adaptive filtering of signals for estimation of the Mahalanobis distance

Abstract

This article focuses on estimating the Mahalanobis distance when there is a dual channel receiving a binary signal. The greatest probability value P _np —adequate signal reception—corresponds to the largest Mahalanobis distance, that is, the optimal tuning of an interference filter. The importance of estimating P _np results from the size of a controlled zone necessary to secure the information being processed: it is found in spurious radiation and regulated by the recommendations of the FSTEK agency, Russia. In contrast to earlier works, proposed algorithms of estimating the value of P _np do not use any information about a learning binary reference sequence. Two types of algorithms are considered for dual channel receiving mode: moving estimates and the Widrow–Hoff stochastic gradient descent.