基于连续波雷达的人体目标微多普勒频率估计方法  

Human micro-Doppler frequency estimation method based on continuous wave radar

作　　者：丁一鹏[1] 柳润金 许雪梅[1] DING Yipeng;LIU Runjin;XU Xuemei(School of Physics and Electronics,Central South University,Changsha 410083,China)

机构地区：[1]中南大学物理与电子学院,湖南长沙410083

出　　处：《中南大学学报（自然科学版）》2022年第4期1273-1280,共8页Journal of Central South University:Science and Technology

基　　金：湖南创新型省份建设专项经费资助项目(2020RC3004);国家自然科学基金资助项目(61501525);湖南省自然科学基金资助项目(2018JJ3680)。

摘　　要：针对连续波多普勒雷达在频谱混叠区域目标信号难以识别以及传统维特比瞬时频率估计算法路径分叉的问题,本文将修正的维特比算法与指数平滑预测技术相结合,根据雷达回波的局部特性动态调整指数平滑法的平滑系数,并定义基于三次动态指数平滑预测的新型惩罚函数,提出一种从连续波雷达回波中估计人体特定散射部位微多普勒频率的新方法。研究结果表明:本文所提方法有效抑制了传统微多普勒频率估计算法在频率模糊区域的路径分叉问题,与基于短时傅里叶变换的峰值检测算法和传统维特比算法相比,采用本文方法所得人体正常行走时左臂微多普勒频率的均方根误差显著降低。新算法所采用的最佳路径搜索方式对传统维特比算法的全平面搜索进行了改进,有效提高了最佳路径的寻找效率。Aiming at the problem that the target signal is hard to identify in frequency ambiguity region of continuous wave radar(CWR),and the path bifurcation problem of traditional Viterbi algorithm in instantaneous frequency estimation,a novel theoretical method to estimate the micro-Doppler frequencies of human specific scattering parts from CWR echo was proposed with a united application of a modified Viterbi algorithm and the exponential smoothing prediction technology.The smoothing coefficient of the exponential smoothing method was adjusted dynamically according to the local characteristics of radar echo,and a new penalty function based on cubic exponential smoothing prediction was defined.The results show that the proposed method effectively suppresses the path bifurcation problem in frequency ambiguity region.Compared with the peak detection algorithm based on short-time Fourier transform and traditional Viterbi algorithm,the root mean square error of left arm micro-Doppler frequency estimation obtained from the proposed method decreases significantly.The employed dynamic search method makes improvement of the traditional whole-plane search and increases the efficiency of searching the optimal paths considerably.

关 键 词：连续波雷达 微多普勒频率 时频分析 维特比算法 指数平滑预测 

分 类 号：TN911[电子电信—通信与信息系统]