The Human Heartbeat And Micro-feature Detection Based On Thz Radar

Terahertz wave is a new undeveloped frequency resource between millimeter wave and infrared. It can easily realize wide signal bandwidth, narrow antenna beamand high precision for its widerspectrum, shorter wavelength, which is suitable for small target detection. Compared with micro-wave and millimeter wave, terahertz wave has a shorter wavelength especially. For the same micro-motion, micro-Doppler characteristic of the target is obviouslyin THz band. It is conductive to micro-motion characteristics observation, feature extraction and motion feature detection of the target. At present, an upsurge of terahertz technology research is seen around the world.As an important crossing frontier, terahertz technology has promising application prospect in through-wall radar, secure communication, security inspection, anti stealth radar, etc, which attracts lots of attention.Because terahertz wave is able to achieve high spatial resolution and isrelatively friendly to human body, terahertz technology has been applied in the biomedical field in the new century, such as detection of cancer, burns imaging and so on.It also shows that terahertz wave is appropriate for human heartbeat and respiration frequency detection.The heartbeat and respiration frequency of human extraction algorithm researchbased on terahertz radar is helpful to translate terahertz technology into specific equipment and promote the application and development of terahertz technology.The Doppler frequency of the moving target is broaden and resolution is improved in THz band because of its unique features, micro-Doppler characteristics of the target in THz band is more sensitive, which make the correlation of target echo and transmitted signal decreased dramatically at the same time, and have a strong impact on the performance of moving target detection, the theory and target detection methods of traditional radar may not be suitable here. The target detection algorithms matched terahertz radar system need to be researched.Meanwhile, the number of cross terms is increase as resolution is improved and micro-Doppler characteristic of the target is more sensitive in THz band. It needs to pay attention to suppress cross terms using the time-frequency analysis method to extract heartbeat and respiration rate. What’s more, the heartbeat and respiration signal belongs to the weak signal relatived to the complex environments. It is necessary to improve the signal-to-noise ratio. The high requirements are put forward for heartbeat and respiration frequency extraction algorithm because of those problems.Aiming at these problems, the research work carried out in this dissertation and the main contributions are as follows:1. The mathematical model of micro-Doppler of radar target is investigated and the conventional micro-movement echo model is analyzed. The performance of some existing time-frequency analysis method is compared through simulation, so are micro-Doppler features of target in THz band and in microwave band. A method of parameter extraction based on Radon transform is presented combining time-frequency analysis method. Compared with conventional methods of parameter estimation, this method has advantages of high-precision and strong anti-noise. The scattering coefficient of each scattering point is estimated for radar targets including multiple scattering points by nonlinear least squares method and the CLEAN algorithm.2. The possible impact of micro-motion of target in THz radar target detection is analyzed, the fact that using conventional algorithms to detect micro-motion target will lead to detection performance declining is testified through theoretical derivation, and the reasons for the decline in detection probability is analyzed. More intense micro-motion of target is, the smaller the correlation between the transmitting and receiving signal is and the lower the detection performance is. Then a joint micro-feature target detection algorithm is proposed based on micro Doppler parameters estimation of target. The detection performance of this method is higher than that of traditional algorithm, which is more suitable for micro target detection. Simulations show that terahertz radar has more advantages than conventional wave band radar and a higher detection probability is obtained under the condition of the same SNR and false alarm probability. This method is less susceptible to micro-Doppler effect and the detection performance is almost unaffected within acceptable range of parameter estimation error.3. The human heartbeat and respiration model, gait motion model and their echo signal are established. The heartbeat and respiration frequency extraction method based on frequency spectrum characteristics is researched. Using over sampling method and low pass filtering, the signal-to-noise ratio of human heartbeat and respiration echo signal is improved. The instantaneous velocity of heartbeat and respiration is obtained by extracting the spectral centroids of the radar return after a time-frequency analysis. The down sampling method is applied to the velocity signal to prevent serious distortion. A second time-frequency analysis is applied to the centroid curve to extract the respiration and heartbeat rates of the individual. The effect of using the ridge line and the centroid curve to approximate heartbeat breathing speed signal on the breathing heartbeat frequency extraction algorithm is simulated. If the resolution of echo spectrum frequency is high, any kind of spectrum optimization method can be adopted, otherwise, the centroid curve should be adopted. Compared to linear time-frequency methods, non-linear time-frequency methods reduce the algorithm steps, improve the efficiency, keep the good time-frequency focusing and have the ability of strong antinoise. Nonlinear frequency extraction algorithm is suitable for weak signal analysis.4. An algorithm that extracts the respiration and heartbeat rates of humans based on empirical mode decomposition technique in terheartz band is presented by improving nonlinear frequency extraction algorithm. Combining smooth pseudo Wigner-Ville distribution and empirical mode decomposition technique, the useful signal frequency components for extracting heartbeat and respiration rates are sifted by appropriate threshold, which further lessens the effects of noise, suppresses the cross-term, and enhances the detection accuracy. The improved algorithm is an effective approach for the detection of respiration and heartbeat signal in a complicated environment. In addition, micro observation experiments by 0.34 THz radar system verify the correctness of the terahertz micro-motion targets echo model in this dissertation.In this dissertation, the micro-Doppler model and the micro-movement echo model of radar target are investigated. A method of parameter extraction based on Radon transform is presented combining time-frequency analysis method. A joint micro-feature target detection algorithm is proposed further. This provides theoretical basis for terahertz radar detection theory and target detection algorithm. Meanwhile, the human heartbeat and respiration model is established. The heartbeat and respiration non-linear time-frequency extraction method based on frequency spectrum characteristics is researched. An Improved algorithm based on empirical mode decomposition technique is presented to further improve the signal-to-noise ratio, which provides an effective method for the analysis of vital signs under the complicated environments.