本論文以數位式信號處理器(DSP)建構一套數位式心脈音量測分析系統,可以提供心音信號的紀錄、播放、雜訊濾除...等即時處理的平台。 此平台採用最先進的TI DSP 之TMS320C6711,是由定點、浮點及多CPU核心的數位信號處理器所組成。TMS320系列亦是針對即時信號處理之應用而設計,同時具備許多優點,如:使用浮點數處理資料、和C語言一樣有親切的架構、高速的性能和具有硬體乘法器。 本研究也應用此平台實現雙麥克風輸入架構,並使用適應性濾波演算法消除背景噪音。此外,也使用時間尺度調整演算法對各種心臟疾病典型之心音信號進行慢速播放處理,使得心肺疾病更易辨認。 將此TMS320C6711 DSK的硬體平台,配合自行開發的人機介面軟體可進行即時信號處理、演算法開發或資料紀錄、重播而建立一套實用的數位式聽診分析量測系統,希望能夠提升心音量測的品質及提供醫師ㄧ個更好的診斷工具,並拓展聽診器在醫學領域方面的應用。
In this thesis, we develop a digital signal processor (DSP)-based platform for heart-sound measurement and analysis. With the DSP kernel, this platform is capable of recording and playing heart-sound signal, as well as digitally filtering background noise. The platform, based on the latest TI DSP chip--TMS320C6711, which features a multiple-CPU architecture, allows it to perform high-speed fixed- and floating-point computation. The TMS320 series of DSP chips are designed for the application of immediate signal processing and has a lot of advantages, such as speedy floating point calculation, built-in hardware multiplexer, and flexible c-like programming language. Based on this DSP platform, we adopt a dual-microphone input in this study, and achieve adaptive noise cancellation with this dual input architecture. A Time Scale Modification algorithm, which can change the replay speed with minimal spectral distortion, is also employed to facilitate the recognition of heart diseases. With self-designed human-machine software, this DSP platform allows real-time signal processing, algorithm developing, data recording and replaying, makes it a practical digital heart-sound measuring and analyzing system. We hope that this system can improve the quality of heart-sound measuring and provide doctors a better diagnostic tool, and further extend the stethoscope application in the medical field.