Research On Feature Extraction Method Of Non-contact Life Signal

Non-contact life signal detection is an emerging technology used for remotely measuring human features of breathing and heartbeat with radar.In recent years,with the improvement of the measurement theory and the reduce of hardware cost,the technology gradually steps into the practical application stage.Under this background,the extraction of two types of life signal characteristics in typical scenes in practical application such as home monitor and medical detection are studied.(1)Real-time extraction method for home monitor.To improve the real-time performance,this method uses the small angle approximate demodulation under the low frequency carrier to obtain the approximate life signal directly.The respiratory rate is extracted real-timely by the time-domain peak seeking method.A fast extraction algorithm based on time division short time window is proposed to overcome the long delay of extracting heart rate in traditional methods.After simulation and experiment,it shows that the single measurement delay is 2s to 5 s,the real-time performance is significantly improved relatively.(2)High-precision extraction method for medical detection.To better accuracy,it obtains radar baseband signal with high phase resolution under high frequency carrier wave.To solve orthogonal imbalances and DC bias on the actual hardware circuit,Ellipse calibration algorithm and Fitting circle algorithm is used to recover the precision motion information.Extended DACM orthogonal demodulation algorithm is proposed to obtain the life signal linear waveform.The simple pendulum experiment is to measure the system's orthogonal unbalance factor and verify that baseband signal preprocessing algorithm is necessary and reliable.The Complete Ensemble Empirical Mode Decomposition(CEEMD)is to separate breath and heartbeat signals,then obtain the heartbeat time-domain waveform with clear interval.Compared with the reference value,the measurement accuracy can reach over 98%.In these two scenarios,the extraction systems of the real-time and high precision are established.After system debugging and the analysis of the performance and error of two systems with the human dithering and respiratory harmonic interference,the corresponding conclusions are drawn,and follow-up work is easier to carry on.