基于微多普勒周期相关性的弹道群目标信号分离

Ballistic Multi-target Signal Separation Based on Micro-doppler Cycle Correlation

针对弹道群目标微多普勒信号分离问题,提出了一种基于微多普勒周期相关性的弹道群目标信号分离算法。对进动目标和旋转目标进行了微动建模,分析了其散射中心在窄带信号下的微多普勒效应,并给出了其参数化表达式。利用STFT对窄带回波进行了时频变换,获取了回波时频图,并采用奇异值分解(SVD)和高斯空间掩模方法分别对时频图进行了消噪和平化处理。再依据散射中心微多普勒瞬时频率的变化规律,采用Viterbi算法对群目标微多普勒曲线进行了提取。最后将提取出的曲线进行傅里叶变换,根据群目标微多普勒周期的相关性,实现了群目标信号的分离。仿真验证了所提方法的有效性和准确性。

Aimed at the problem that ballistic multi-target micro-Doppler signal separation,an algorithm based on micro-Doppler cycle correlation is proposed for separating the ballistic multi-aspect signal.Cone-shaped precession model and rotating target model are established.After that,targets’scattering centers are analyzed in the narrowband signal and their parametric expressions are given.Using Short-Time Fourier transformed scattering centers echo into time-frequency figure.Then,applying SVD method to eliminate time-frequency figure’s noise and using Gaussian space mask method to smooth time-frequency curve.Moreover,Viterbi algorithm is used to extract multi-target micro-Doppler curves based on the variation law of micro-Doppler instantaneous frequency.Finally,the extracted curves are transformed by FFT,and multi-target signal are separated based on micro-Doppler cycle correlation.Simulations verify the effectives and accuracy of the proposed method.