| Non-contact vital sign detection method is a method that uses non-contact sensors to collect sign signals and detect human cardiopulmonary activities.Compared with traditional contact sign detection method,it can avoid interfering with human feelings and affecting the quality of life.The study of millimeter-wave radar as a non-contact sign detection sensor for heart and lung information extraction has attracted extensive attention.This paper focuses on the systematic research on the process and technology of millimeter-wave radar extraction and separation of vital signs.The main work and achievements are summarized as follows:Firstly,sign detection scheme based on FMCW millimeter wave radar is designed and the sign signal model is analyzed.According to the requirement of physical signs detection,the selection of radar hardware was determined,and the working principle of FMCW radar physical signs detection was analyzed.According to the principle of thoracic movement and signal analysis,a sign signal model based on cardiopulmonary activity was established,and a heartbeat model based on Gaussian pulse sequence and an exponential attenuation sinusoidal pulse sequence were studied.Secondly,the preprocessing process of the sign signal based on millimeter wave radar is analyzed.The phase signal extraction method was studied,the clutter interference in phase difference signal was analyzed,the clutter removal method based on exponential weighted moving average filter was studied,and the body motion detection method based on double threshold energy threshold was studied.The feasibility of the body motion detection method was verified through experiments.Finally,the cardiopulmonary information detection method based on radar sensor is researched and verified.This study mainly focuses on the requirements of monitoring physiological parameters after signal acquisition,and emphasizes the accuracy,low cost and reference of the detection method.In order to extract the changes of respiratory motion,the wavelet entropy based on double threshold was firstly used to identify apnea,and then the improved ensemble empirical mode decomposition method was used to separate and identify the respiratory waveform.In the aspect of heartbeat information detection,the method of estimating heart rate based on the high frequency component of heartbeat was studied,and the heartbeat waveform was reconstructed by using the estimated heart rate,integrated empirical mode component and the proposed heartbeat model.The experimental results show that the difference respiratory signals and heartbeat signals can be separated accurately under the sitting and static state.The time-domain information error of HRV parameters and ECG calculation results is less than 0.005,and the frequency domain information error is less than 0.1,and the extracted heartbeat interval is in good agreement with the ECG separated heartbeat interval.And it is accurately to detect the occurrence of apnea or body movement disturbance time period. |
