Cyclic system reconfiguration for adaptive separation of lung and heart sounds

Abstract

This paper introduces an improved methodology of signal separation for removing noise and separating heart and lung sounds. Unlike traditional noise cancellation and signal separation methods that rely on frequency band separation or statistical independence to achieve signal separation, this methodology utilizes the unique feature of time-split stages in breath and cardiac cycles to simplify the blind source separation problem and enhance algorithm performance. By employing a multi-sensor system, the method performs time-split channel identification, adaptive signal separation, and noise cancellation, with recursion from cycle to cycle. Since no frequency separation or statistical independence is required, this method can provide a robust and reliable capability of signal extraction for enhancing accuracy and reliability of medical diagnosis.