信号波形已知条件下多目标直接定位性能分析及其改进算法

相比传统"测向+位置估计"两步定位体制,以Weiss和Amar提出的目标直接定位技术具有精度高、分辨能力强等优点,并且便于利用信号波形信息.该文首先针对Weiss和Amar提出的信号波形已知条件下的多目标直接定位算法（称为Weiss-Amar算法）进行理论性能分析,证明了其参数估计的渐近一致性,推导了其参数估计的渐近方差,并指出当不同目标信号波形相关时会对算法性能产生较大影响.针对这一缺点,文中提出了信号波形已知条件下的改进型多目标直接定位算法,与Weiss-Amar算法不同的是,改进算法对全部目标的位置参数进行协同估计,并且通过交替投影迭代进行数值优化,能显著提高对相关信号的定位精度.最后文中推导了信号波形已知条件下多目标位置直接估计方差的克拉美罗界.仿真实验验证了文中理论分析的有效性和改进算法的优越性.

A Performance Analysis and Improvement Algorithm for Multiple-Source Direct Localization with Known Signal Waveforms

Compared to the conventional two-step (including direction finding and position estimation) localization mechanism,the direct position determination (DPD) approach presented by Weiss and Amar has lower estimation errors and stronger resolution capability.Moreover,this technique provides a more convenient way to incorporate the signal waveform information.In this paper,the theoretical performance of the DPD algorithm presented by Weiss and Amar (called WeissAmar algorithm) is studied analytically.The asymptotical consistency is proved and the closed-form expression of the estimation variance is derived based on the error perturbation analysis method.It is also noted that the localization accuracy of this algorithm will decrease if the signals are correlated with each other.In order to solve this problem,an improved DPD algorithrn is proposed for the case of known waveforms.In the new algorithm,the position parameters of all targets are considered as a whole to be estimated by using the alternating projection iteration.The performance improvement compared to the Weiss-Amar algorithm is remarkable for correlated signals.Furthermore,the Cramér-Rao bound (CRB) on the direct position estimation in presence of known waveforms is also deduced.Simulation results corroborate the effectiveness of the theoretical analysis and the advantage of the proposed algorithm.