Research And Implementation Of Long-Term Vital Sign Monitoring Based On Ultra-Wideband Radar

Vital signs are important indicatiors of the existence and quality of life activities.Monitoring of vital signs for the human body is an important basis for disease prevention and clinical diagnosis.Ultra-Wideband(UWB)radar,with low power consumption and high resolution,provides a non-contact monitoring measurement,and effectively avoids light interference and privacy violations.UWB radar has been widely used in many vital sign monitoring systems.However,the existing vital sign monitoring systems do not discuss the simultaneous monitoring of vital signs for multiple targets.And there is still insufficient research on how to remove the interference caused by target random body motion.In order to address the above two problems,this paper proposes in-car multiple targets vital sign monitoring using location-based Variational Mode Decomposition(VMD)algorithm and waveform joint optimization of body movement compensation based on sinusoidal data fitting and mapping for moving target vital sign monitoring.Multiple targets vital sign monitoring algorithm under complex strong reflection environment is proposed in this paper,which solves the problem that it is difficult to distinguish the vital sign signals of multiple targets under the condition of metal multipath interference and multi-target multipath aliasing interference.VMD is performed on each signal interval,which is divided from the received radar signals based on the prior knowledge.The vital sign signals from the same human body have better correlation in the time domain.According to this feature,when extracting vital signs from targets farther away from the radar,the correlation coefficients between extracted vital signals in different locations are compared and the signal with the highest correlation is removed as the superposed signal to achieve multiple targets interference removal.And joint waveform optimization against body movement interference is proposed for moving target vital sign monitoring.In the complex environment of strong reflection inside the car,the unavoidable driving motion causes the periodicity of the signal to be destroyed,make it a challenge to recover the highly variable cardiopulmonary waveforms.Two algorithms including the sinusoidal data fitting based and mapping based body movement compensation algorithms are implemented to resist interference from motion.The features of the waveform in the time and frequency domains are extracted,and these signals are given weights in combination with the random forest.The Kalman filter is used for waveform fusion to realize joint optimization of waveforms to obtain the reliable vital sign waveforms.In order to verify the effectiveness and practicality of the above algorithms,an actual test environment is set up and long-term radar data is collected.On this data set,the proposed algorithm is compared with the existing algorithms in terms of performance.The experimental results show that the algorithm proposed in this paper completes the long-term vital signs monitoring in a closed space(in car),and the accuracy can reach 92.66%and 88.41%under static and moving states,respectively.