Monitoring the propagation of lung sounds using electronic stethoscope arrays

This thesis presents the design and prototype testing of a novel medical instrument designed to measure changes in to acoustic transmission properties of lung tissue. Since tissue acoustic transmission is largely determined by the distribution of lung fluid and lung tissue density, this instrument has potential applications for monitoring and diagnosis of patients with such obstructive lung diseases which are associated with accumulation of lung fluid and collapse of lung tissue. The apparatus consists of an array of 4 electronic stethoscopes linked together via a fully adjustable harness. A White Gaussian Noise (WGN) input sound is injected into the mouth via a modified speaker and measured on the surface of the chest using the array of stethoscopes. Data were analysed using the Normalized Least Mean Squares (NLMS) adaptive filtering algorithm to develop a transfer function based on the propagation characteristics of the injected signal. This transfer function is then analysed to determine the frequency response and the propagation delay at each stethoscope. The system was calibrated to account for delays in the signal acquisition equipment and verified using a chest phantom model. System non-linearites were analysed and determined to be sufficiently small to justify the linear model. Phantom test results show that as the volume of fluid in the lungs increases, the sound propagation delay decreases. In-vivo results were measured on healthy volunteers and show comparable results to the lung phantom with no volume of water injected and that the instrument can detect sound propagation delay variations with changes in posture. Based on these results, this instrument is able to measure parameters of the lungs including propagation delay and frequency/impulse responses, which show useful correspondence to known physiological changes.