Research On Signal Processing Method Of Human Micro-motion Information Extraction Based On Doppler Radar

In recent years,natural disasters have occurred frequently,rescuing survivors have become the most urgent task after the disaster.However,traditional vital signal detection techniques have many disadvantages and cannot be applied to complex on-site environments.Under such circumstances,we urgently need a detection device to deal with the complex environments and accurately locate the target is urgently needed.Electromagnetic waves having strong penetration and anti-interference capability,therefore,radar can be used to realize the non-contact life detection.The human respiration and heartbeat are one of the most prominent features of life.Radar non-contact life detection technology which detects the Doppler frequency shift caused by the human respiration and heartbeat can be used to detect vital signals.In this thesis,we use continuous wave as the transmitting signal of radar.Based on continuous wave radar,signal models of human micro-motion and radar echo are established and two human target detection algorithms are proposed.The main research work of this thesis is as follows:(1)The principle of Doppler radar is introduced.According to the periodicity of human respiration and heartbeat,the Doppler radar micro-motion signal model are established.Under the two radar systems of the single-frequency continuous wave and the step-frequency continuous wave,the radar echo signal models after target reflection is established.(2)In the case of the single-frequency CW radar system,this thesis proposes the complex continuous basis pursuit(CCBP)based vital signal detection algorithm.The proposed CCBP algorithm can extract the vital signal information by using the polar interpolation and the convex optimization problem solving technique,which can accurately estimate the frequency of relatively weak heartbeat signal.However,the traditional Multiple Signal Classification(MUSIC)algorithm fails to detect the weak heartbeat signal at low signal-to-noise ratio.Simulation experiment results have demonstrated that the proposed CCBP algorithm reduces Root Mean Square Error and improves the probability of successful estimation,which is more suitable for the vital sign detection than the MUSIC algorithm.(3)In the case of the step-frequency continuous wave radar system,in order to reduce the amount of the data samples,the simultaneous orthogonal matching pursuit(SOMP)reconstruction algorithm is proposed to extract the range profile of the human target under the Compressive Sensing framework.The orthogonal matching pursuit(OMP)algorithm is then used to extract the human respiratory and heartbeat signal frequencies.The simulation experiments verify that compared with the traditional OMP algorithm in both the frequency dimension and the slow time dimension,the proposed algorithm is more suitable for determining the human distance and extracting the respiratory and heart micro-motion signal frequency.