Human Body Posture,Breathing And Heartbeat Detection Based On Millimeter Wave Radar

Human body posture and breathing and heartbeat detection have important application value in post-disaster rescue,patient monitoring,and urban warfare.Traditional human posture detection uses optical equipment to collect data,and the detection results are easily affected by factors such as light and environment;traditional respiratory and heartbeat detection mostly uses contact devices,which have problems such as easy cross-infection and low efficiency in the detection process.Respiratory,heartbeat and motion posture detection based on millimeter wave radar is not only unaffected by changes in illumination,but because it is non-contact detection,it has a wide range of application scenarios.This topic uses millimeter-wave band AWR1642 linear frequency modulation continuous wave radar to collect human breathing,heartbeat,and human motion posture data,and perform preprocessing,algorithm simulation,and actual measurement data processing and analysis on echo data.The specific content is as follows:1.Introduced the AWR1642 millimeter wave radar and the basic principles of respiratory and heartbeat measurement,and described the radar parameter settings,echo data format and the preprocessing of respiratory and heartbeat data.2.Perform distance gate selection,phase extraction,down-sampling and other processing on the respiratory and heartbeat echo data.On this basis,the heartbeat breathing frequency estimation based on the energy center of gravity correction method is introduced.The results show that the energy center of gravity correction method can improve the spectral line.Misaligned data set detection results.For the error caused by the non-energy center of gravity problem,a respiratory and heartbeat data reconstruction based on the empirical mode decomposition method is proposed.In the theoretical simulation and actual data processing and analysis,this method improves the accuracy of the respiratory and heartbeat detection.3.On the basis of the empirical mode decomposition signal reconstruction method,the instantaneous frequency of breathing and heartbeat is estimated,and the instantaneous frequency estimation method of joint peak and valley peak finding method is proposed,and the Hilbert Huang transform time-frequency analysis method is used to analyze the respiratory rate.The heartbeat data was processed and analyzed,and the results proved the accuracy and stability of the test results.4.In the human body gesture recognition,a motion gesture feature extraction method based on dynamic Doppler spectrum is proposed.First,on the basis of echo data preprocessing,the distance Doppler spectrum of moving target is obtained through fast-time and slow-time two-dimensional fast Fourier transform.Based on the distance Doppler spectrum,a sliding window target area detection algorithm is proposed.Perform target detection in the target area.Based on the energy-weighted representative point extraction method for the detected target points,the velocity,distance and energy representative points of the dynamic Doppler spectrum are calculated respectively,and the motion characteristics of the dynamic Doppler spectrum are extracted on this basis.Finally,the support vector machine method is used to learn,identify,and analyze the human body movement posture feature data.The experimental results verify that the algorithm has high accuracy and robustness.