一种基于最大似然估计的合作目标多维参数跟踪算法

Maximum Likelihood Estimation Based Algorithm for Tracking Cooperative Target

空间交会对接微波雷达采用基于延迟锁定环(DDLL)、锁频环(FLL)和锁相环(PLL)的算法处理合作目标转发的直接序列扩频信号,获得入射信号的时延、频率及波达角(DOA)估计。针对当前基于DDLL,FLL和PLL(DFP)的算法没有充分利用接收信号有效信息的问题,该文提出一种基于极大似然估计(MLE)的低代价闭环跟踪(MLBT)算法。该算法利用代价函数的梯度正比于参数误差的特性,设计了参数误差鉴别器。在此基础上给出了相应的扩频信号多参数跟踪环路。分析并验证了鉴别器的方差特性,从而给出MLBT算法的参数估计均方根误差(RMSEs)下限。仿真实验验证了对MLBT算法参数估计均方根误差下限分析的正确性。此外仿真结果表明,MLBT算法相比DFP算法有限地增加运算量,同时获得了更优的参数估计性能。

The scheme which is based on the Digital Delay Locked Loop （DDLL）, Frequency Locked Loop （FLL）, and Phase Locked Loop （PLL） is implemented in the microwave radar for spatial rendezvous and docking, and the delay, frequency and Direction Of Arrival （DOA） estimations of the incident direct-sequence spread spectrum signal transmitted by cooperative target are obtained. Yet the DDLL, FLL, and PLL （DFP） based scheme has not made full use of the received signal. For this reason, a novel Maximum Likelihood Estimation （MLE） Based Tracking （MLBT） algorithm with a low computational burden is proposed. The feature that the gradients of cost function are proportional to parameter errors is employed to design discriminators of parameter errors. Then three tracking loops are set up to provide the parameter estimations. In the following section, the variance characteristics of discriminators are investigated, and the low bounds of Root Mean Square Errors （RMSEs） of parameter estimations are given for the MLBT algorithm. Finally, the simulations and computational efficiency analysis are provided. The low bounds of Root Mean Square Errors （RMSEs） of parameter estimations are verified. Additionally, it is also shown that the MLBT algorithm achieves better performances in terms of estimators accuracy than those of the DFP based scheme with a limited increase in computational burden.