Studies On Heart Rate Detecting Methods And Devicesfor Sleep Analysis

Recent research shows that heart rate monitor can be used to evaluate sleep quality. It offers a simple way to analyze dormancy. The thesis focuses on the detecting devices of three kinds of biomedical signals including electrocardiogram (ECG), phonocardiogram (PCG) and pulse wave. All these signals contain the heart rate information.The design of the multi-channel bioelectric signal acquisition module is based on Micro-Controller-Unit (MCU) which can measure ECG and detect heart rate during sleep. The system uses battery as power supply. Its characteristics are low power consumption, small volume, long working duration, high sampling rate, etc. High resolution A/D converter and 8051 MCU are combined as key hardware structure which even can be used as electroencephalogram (EEG) signal acquisition.Phonocardiogram (PCG) detection system based on Digital Signal Processor (DSP), not only is applied for heart rate detection, but also can be used as an integrated digital electronic auscultation system. According to the function it can be regarded as DSP subsystem and Matlab GUI subsystem. In the former single-band real-time modulation for PCG signal is realized by Hilbert transformation on the DSP platform which enhances acoustical effect of heart sound. The latter subsystem controls PCG signal acquisition through the computer's audio card under Matlab graphic user interface (GUI).The pulse acquisition system based on SOC (System On Chip) can be applied both in heart rate and blood oxygen saturation detection during sleep. The noninvasive optical monitoring theory of pulse oxygen, the system's hardware structure and software design flow are discussed in detail. The system is mainly composed of light driver, photoelectric transformation and MSC1210 which is a mixed signal processing chip that integrates 24bits A/D converter and enhanced 8051 core.Finally, for the digital signal preprocess, a method which combines high pass filter with fraction delay filter is proposed in this thesis. The design of FIR comb filter becomes possible even when the sample rate (fs) is not a multiple of the notch frequency (fp). The optimized filter structure has also been described. The filter designed by this approach has linear phase in pass band and good notch performance. Promising result has been achieved when the filter is applied to ECG signal contaminated by baseline wandering, power line interference and its harmonic components.