Study On Human Target Micro-Doppler Frequency Estimation Algorithm Based On Continuous Wave Radar

Continuous wave radar plays an important role in practical application fields such as target micro motion sensing,positioning and tracking and state recognition.Based on the dual frequency continuous wave radar system,this paper studies the micro-Doppler estimation algorithm of each scattering part of the human body,and completes the real-time tracking and positioning of the target according to the estimated micro-Doppler frequency results.Aiming at the component Doppler frequency aliasing caused by the interaction of human joints,an improved Viterbi algorithm based on exponential smoothing prediction theory is proposed.(1)Aiming at the common path bifurcation problem of traditional Viterbi algorithm in the process of target micro-Doppler frequency estimation,the algorithm constructs a multi-dimensional penalty function to complete the automatic selection of path direction and the adaptive adjustment of path parameters combined with Hough transform,so as to realize the optimal path fitting of scattering characteristics of different parts of human body and improve the estimation accuracy of micro-Doppler frequency.(2)By optimizing the search mode of multiple candidate paths,the algorithm greatly improves the search efficiency of the best path.The experimental results show that the algorithm can achieve excellent detection performance in the harsh detection scene with complex target trajectory,large echo dynamic range and high real-time demand for signal processing,and is suitable for the high-precision estimation of human target micro-Doppler frequency.Aiming at the characteristics of large dynamic range,strong nonstationary and serious instantaneous frequency aliasing of scattering components in various parts of human body,a micro-Doppler frequency estimation algorithm of human body position based on ARMA model is proposed and applied to real-time tracking and positioning of targets.(1)Compared with other scattering parts,the scattering component of human trunk has the characteristics of more stable energy,narrower dynamic range of micro-Doppler frequency and more obvious linearization trend.Therefore,an algorithm is proposed to estimate the micro-Doppler frequency of the human trunk.(2)Combined with Doppler frequency integration processing,the realtime position information of the target is obtained.The proposed algorithm can effectively solve the common trajectory mismatch problem in the traditional positioning algorithm.Even in the area where the multi-target time-frequency curves overlap each other,it can accurately extract and estimate the micro-Doppler frequency of the target trunk,and then complete the real-time positioning and tracking of the target.