Double-microphone self-adaptive filtering algorithm used for collecting body sound signals and application

Abstract

The invention provides a double-microphone self-adaptive filtering algorithm used for collecting body sound signals. The double-microphone self-adaptive filtering algorithm is characterized in that one main microphone and one auxiliary microphone are at least adopted for collecting signals; the main microphone is used for collecting signals with noise, and the auxiliary microphone is used for collecting environmental noise; identical high-pass filtering processing is conducted on the signals collected by the main microphone and signals collected by the auxiliary microphone; normalized minimummean square algorithm is adopted for the main microphone signals and the auxiliary microphone signals after high-pass filtering processing to calculate a weight value of a self-adaptation filter device and calculate error signals to filter away the environmental noise in the main microphone signals; primary low-pass filtering processing is conducted on the error signals to restore body sound signals, so that the body sound signals output by the self-adaptive filtering algorithm are obtained. The algorithm can conduct rapid converging on the weight value of the filter device and avoid signal distortion, and rapidly and reliably restrain interference of the environmental noise.

Description

technical field [0001] The invention relates to the technical field of medical measurement and signal processing, in particular to a dual-microphone adaptive filtering algorithm and application for collecting body sound signals. Background technique [0002] Remote auscultation enables users to enjoy telemedicine services without leaving home, making it possible to see a doctor efficiently anytime, anywhere, and greatly reducing the cost of follow-up for chronic disease patients. However, remote auscultation has high requirements on the anti-noise capability of the auscultation system: weak body sound signals are easily disturbed by environmental noise, and during the process of remote auscultation, the doctor does not know the situation of the patient's environment, and it is difficult to judge the abnormality heard. Whether the sound is the murmur of the patient's body sound or the ambient noise is prone to misdiagnosis. Therefore, the remote auscultation system must take...

AI Technical Summary

Problems solved by technology

[0006] In the application of the normalized least mean square algorithm, the reasonable value of the adjustment factor η is very important: the filter weight iteration depends on the adjustment factor ηW(k+1,i)=W(k,i)+η(d (k)-y(k)) / ε(k); wherein, d(k)=s(k)+n(k), s(k) and n(k) are body sound signals at the kth moment respectively and environmental noise; while the adaptive filter output is: e(k)=d(k)-y(k); if the adjustment factor is too small, the convergence will be slow, and the purpose of suppressing environmental noise cannot be achieved for a long time; and if the adjustment factor If it is too large, it will easily cause the filter weight W(k+1,i) to diverge, and the adaptive filter will fail

[0007] If the traditional normalized least mean square algorithm is applied to auscultation filtering, when the amplitude of the body sound signal s(k) is much larger than the amplitude of the environmental noise n(k), the first and second heart sound amplitudes are very large , it is very easy to cause the output distortion due to the overadjustment of the adaptive filter parameters due to the excessive value of the adjustment factor η

If a small adjustment factor η is selected to reduce the degree of signal distortion, the filter weights will converge too slowly and lose the value of practical application.

The contradiction between signal fidelity and fast convergence is difficult to overcome in the general normalized least mean square algorithm

[0008] Therefore, the dual-microphone adaptive filtering algorithm still cannot be directly applied to electronic auscultation.

Images