基于DML-ESPRIT算法的多维参数快速联合估计

将确定性最大似然估计准则的多维参数估计能力与ESPRIT算法的高时效性有机结合,提出了一种二维DOA-功率-频率快速联合估计方法--DML-ESPRIT算法。利用双L阵列的空间特性,通过引入空间锥角,将多维空间搜索问题转换为一维角度估计,并基于确定性最大似然估计准则推导得到了多信源的空间锥角、功率和频率联合估计的数学模型;然后在对子阵进行扩展的基础上,利用TLS-ESPRIT算法对模型进行求解,避免了谱峰搜索问题,实现了多维参数的快速联合估计。实验结果表明,DML-ESPRIT算法在保持高估计精度的同时运行耗时约35 ms,具有较好的工程应用前景。

Multidimensional Parameter Estimation Based on DML-ESPRIT Algorithm Fast Joint

By organically integrating the joint multidimensional parameters estimation ability of deterministic maximum likelihood (DML) estimation criterion and the computation efficiency of estimating signal parameters via rotational invariance techniques (ESPRIT),a novel joint two-dimensional DOA (2-D DOA),power and frequency fast estimation algorithm named DML-ESPRIT is proposed.Firstly,based on the special characteristics of double L-shaped array,we introduce the space cone angle to represent the source's 2-D DOA.This operation successfully converts the multidimensional space searching problem into the separated one-dimensional angle estimation.Under the DML estimation criterion,the joint estimation model of parameters such as space cone angle,signal power and frequency is established.Then TLS-ESPRIT method which could avoid the spectrum peak search is used to solve the model.On the premise of keeping estimation precision,the time consumption of DML-ESPRIT algorithm is only about 35 ms on the average,thus good prospects for engineering applications.