Heart Rate And Respiratory Rate Estimation Based On Bayesian Statistical Theory And MIMO Radar

With the development of society,people pay more and more attention to their own health,so it is particularly necessary to develop a technology that can easily collect human health data.As an important branch of Non-contact monitoring,radar-based vital signs measuring technology emerges at the right moment.Multiple-Input Multiple-Output（MIMO）radar can achieve superior performance than traditional radar.Therefore,the analysis on heartbeat and respiration monitoring methods based on MIMO radar is proposed to improve the estimation performance of vital sign parameters such as heart rate and respiration rate.It also provides theoretical reference for the optimization design of the corresponding radar system.Radar detects chest fluctuation caused by cardiopulmonary movement to monitor vital signs.As a traditional method,time-frequency analysis method estimates heart rate and respiration rate by utilizing frequency domain characteristics of signals.MIMO radar signal model containing vital signs information is developed in this thesis.The applications of two classical time-frequency analysis methods,Complex Signal Demodulation（CSD）method and Arctangent Demodulation（AD）method,are extended in MIMO conditions.Through theoretical derivation of spectrum characteristic and numerical simulation analysis,it is confirmed that both methods can make relatively accurate estimation of respiration rate.However,due to the existence of the harmonic interference problem,AD method can estimate the heart rate when the condition is sufficient,while the CSD method cannot estimate heart rate stably.In additional,both methods are significantly affected by SNR,observation duration,and spectral accuracy,which limit the further improvement of estimation accuracy.In order to further improve performance,Maximum Likelihood（ML）estimation method based on bayesian statistical theory is adopted.The ML estimate and CramerRao Bound（CRB）are derived and the correctness of theoretical derivations is proved by simulation experiments.The results indicate that CRB can be used as a reliable metric to evaluate the performance of the estimation of heart rate and respiration rate.The performance of CSD,AD,and ML estimation methods is compared under different pulse numbers.This thesis also utilizes CRB to analyze the causes of the difficulties in heart rate estimation and the effects of radar parameters,such as carrier frequency,signal timewidth,orthogonality of transmitting signal and number of antennas,on the estimation performance.Finally,different antenna placement strategies are simulated,and the results show that optimized antenna arrangement can significantly improve the estimation accuracy of heart rate and respiration rate.