Publication de 6 numéros par an
ISSN Imprimer: 2152-5102
ISSN En ligne: 2152-5110
Indexed in
Statistical Estimation of the Acoustic Parameters of a Human Respiratory Tract
RÉSUMÉ
The paper addresses a statistical approach for handling the problem of estimation of the acoustic parameters of a human respiratory tract. Synthesis and analysis of the estimation algorithms under complicated interference conditions accounting the presence of the heart's tones are performed. It is shown that the interference of the heart's tones can be suppressed through the use of compensating channels formed with the view of the correlation characteristics of interference in the frequency or time domains. The basic procedures of input signal processing and the ways of acquiring the data required for their execution are determined. Potential efficiency of the algorithms for a number of models of the space correlation functions of the breath noise sources is examined. It is noted that, at large values of the ratio "signal/interference", the variance of estimations of acoustic parameters to be measured is of low sensitivity to modification of the form of the space correlation function of the breath noise sources. The results of experimental research aimed at estimation of the sound speed in the lung parenchyma are presented. The algorithm developed in the paper is shown to provide the unbiased estimate of the sound speed, being in concord with the known experimental data for the lung parenchyma.