Design And Implementation Of Respiratory And Heartbeat Detection System Based On Millimeter Wave Radar

Respiratory and heartbeat signals are important signals of human life characteristics.At present,photoelectric volumetric pulse wave instrument(PPG)and electrocardiogram monitor(ECG)are the main detection methods for contact detection,which have great limitations.In order to solve this problem of medical monitoring,non-contact breathing and heartbeat monitoring techniques have been developed.Millimeter wave radar has the advantages of narrow beam,wide band,small volume,resolution and no harm to human body,so it has a broad research prospect in medical and health care fields.Based on pulse coherent millimeter wave radar,the respiratory heartbeat detection system is designed and implemented in this paper.Firstly,the signal characteristics and principle of pulsed coherent millimeter-wave radar are introduced,and the radar information used in the project is introduced.Then,starting from the data stream,design the data transmission and data structure of the radar data acquisition and transmission module,and then carry out the radar data transmission between the server and the client through Socket connection.Finally,the original data of the first step is obtained by phase processing,and then the respiratory signal and heartbeat signal are separated and extracted by filtering algorithm.Then the breath and heartbeat frequencies of the separated phase data were detected,and the breath and heartbeat signals were detected and analyzed by using short-time Fourier transform(STFT),principal component analysis(PCA),Morlet wavelet transform(MORT)and multiple signal classification algorithm(MUSIC).Short-time Fourier transform(STFT)needs to adopt appropriate window to balance the contradiction between time resolution and frequency resolution.Principal component analysis(PCA)can analyze the change of signal from the perspective of time domain.Morlet wavelet transform can provide instantaneous frequency change information and accurate detection results.Music algorithm can obtain high resolution frequency information through spectrum estimation,but the computational complexity is high.Through the comparative test analysis of single sitting,the accuracy of the detection value obtained by the Morlet wavelet algorithm reaches more than 90%,and it can provide abundant instantaneous frequency information at the same time.In addition,the detection time is less,and the real-time performance is high.Therefore,Morlet wavelet transform is a better choice for the detection of breathing and heartbeat signals.Finally,Morlet wavelet transform and multiple signal classification algorithm are combined to analyze the detection possibility of three kinds of complex scenes.The experiment shows that the detection is still reliable under single person and multiple angles.In the case of two-person detection,the respiration and heartbeat information of each person still exists,but the signal separation and extraction detection of each person needs further analysis and improvement.Morlet wavelet transform algorithm can detect breathing and heartbeat signals under single-person sitting and shaking scenarios,but it is easy to be covered by motion signals and has poor reliability,which requires further improvement.