Research On Home Non-contact Vital Sign Monitoring System Based On Doppler Radar

As the pace of life continues to accelerate,sudden deaths caused by sudden heart disease are common,and the best treatment time is often only a few minutes.To deal with such sudden diseases,the best solution is to monitor physical signs such as heartbeat and breathing,and give targeted early warning.There are many monitoring instruments on the market,but most of them are not suitable for long-term monitoring.Contact measurement will also affect the quality of life.In this thesis,a Doppler radar vital sign monitoring system for homes is designed to monitor human heartbeat,breathing and sleep status.Aiming at the problem of noise and clutter interference in radar signals,analyze the influence of slight human body movement on respiratory and heartbeat signals,and judge whether human body movement occurs within the selected time window.Improve the accuracy of the signal under low signal-to-noise ratio through the method of autocorrelation detection,and finally improve the anti-noise ability of the system.Design a digital filter combined with short-time Fourier transform to perform time-frequency analysis on the signal to extract the heartbeat and respiratory frequency information.In order to achieve accurate staging of human sleep status,and solve the problem of low accuracy in extracting HRV feature parameters in a non-contact manner,an HRV analysis algorithm based on BPFP is proposed.Based on the estimation of the heartbeat frequency,the IBI information is extracted from the zero-crossing point detection of the output signal of the narrow-band filter bank to calculate the HRV feature,combined with the characteristic parameters such as heart rate,breathing and body movement,the method based on the support vector machine is selected for sleep staging.In terms of performance optimization of support vector machines,the traditional particle swarm algorithm has insufficient precision and falls into local optimality.According to the fitness of the candidate solution space,the inertia weight is dynamically selected,and the adaptive mutation operator is added to increase the diversity and ensure the overall optimization ability of the particle swarm algorithm.Experimental data shows that the SVM model under the improved algorithm has higher classification accuracy than GA,PSO,and BP algorithms,and can achieve a staging accuracy of 91.24%.Finally,a vital signs monitoring system based on IWR1642 millimeter wave Doppler radar was established,and the heartbeat and breathing information detection was completed through the radar front end and signal processing module.Through experimental comparison,the accuracy of the system monitoring is verified,and it can provide effective data reference for personal health assessment.