针对稀疏多子带信号多子带综合时差估计难,高分辨时差估计计算量大的问题,提出了一种基于线性调频Z变换(CZT)的稀疏多子带信号融合的快速高分辨时差估计算法。该方法通过CZT对各子带信号的互谱进行高分辨的谱分析得到各子带信号的高分辨互相关函数,通过分析各子带互相关函数的相位关系实现各子带互相关函数的相参积累。相比传统广义互相关算法,该方法在实现多子带相参积累的同时,大大降低了计算量。仿真结果表明:提出的基于CZT的多子带融合快速时差估计方法能精确地估计时差,相比单子带信号的时差估计性能可以获得明显提升。

Since there exists some problems in the time difference of arrival (TDOA) estimation of multi-band signals, such as the coherent accumulation of multi-band and the computation cost of high resolution TDOA estimation, a chirp-Z transform (CZT) based TDOA estimation method is proposed for multi-band signals, which can get high resolution and accurate TDOA estimation quickly through coherent accumulation of multi-band cross correlation. The proposed method utilizes the CZT to analyze the spectrum of each band signal's cross-spectrum to get high resolution cross correlation of each band. Then the phase relations of cross correlations between each band are analyzed, so the cross correlations of each band can be accumulated coherently to get a fused cross correlation. Compared to the traditional generalized cross correlation method, the proposed method also achieves coherent accumulation of multi-band and has much less computation. The results of computer simulations demonstrate that the proposed methods can get accurate TDOA estimation results, and the performance has been obviously improved compared to the estimation results of single band signal.

TN957.51