Research On ECG Signal Processing Based On Body Sensor Networks

Cardiovascular diseases are a major kind of diseases that endanger human health, of which the diagnosis is of the most significance. Currently, electrocardiosignal monitoring is usually adopted to diagnose the cardiovascular diseases clinically. Resting electrocardiograms (ECG) are widely used in conventional monitoring. However, it is able to return correct monitoring results only on the occurrence of morbidity, which severely limits its application. On account of this limitation, long-term monitoring is made possible by the presence of dynamic ECG. But, body movement of patients and environment has bad effects on the electrocardiosignal detection. In order to overcome this weakness, the thesis proposes a method based on body sensor network (BSN) for better electrocardiosignal processing, which introduces the triaxial accelerations to separate the real signals and noise signals for the more precise electrocardiosignal detection.With the help of BSN, a context-aware electrocardiosignal detection method is designed in the thesis to determine patients’ movement state according to their accelerations and to choose the appropriate algorithm for the signal detection. As for the state of rest, a filter-threshold algorithm is used for signal detection. This algorithm eliminates the power-line interference and the baseline drift with a60Hz wave trapper and a median-filter-based curve-fitting method respectively. Then the peaks of the R-wave are detected with an self-adaptive threshold method, and the Q-wave, the S-wave, the P-wave and the T-wave are detected through local maximum value searching and the time window method. The validity of the algorithm is verified using from both the data from MIT-BIH database and the data collected from real resting patients. The experimental results indicate that the algorithm is able to correctly detect the feature points of electrocardiosignals. When the patients are in the motion state, the thesis adopts the wavelet transform method to eliminate the myoelectrical interference, the power-line interference and the baseline drift. Then the wavelet transform method is used to detect the peaks of the R-wave, after which the Q-wave, the S-wave, the P-wave and the T-wave are detected through local maximum value searching and the time window method. The algorithm is verified with the data from MIT-BIH database and the data collected from real moving patients. The experimental results show that the algorithm is capable of detecting the feature points of electrocardiosignals correctly. On the ground of the approaches mentioned above, the separation of the real electrocardiosignals and the noise signals and the precise detection of the electrocardiosignals are achieved.