Abstract: Based on 2-D virtual ESPRIT algorithm，a high performance improved method is presented.The new method smooths the subarray data by subarray data direct spatial smoothing.And then virtual array transformation technology is used to realize array transformation.The 2-D angle parameters are computed by DOA matrix method.Simulation results indicate that compared with spatial smoothing 2-D virtual ESPRIT algorithm and spatial smoothing DOA matrix method，the method has better estimation performance for coherent signal sources in low SNR environments，and it can estimate more independent signal sources than the number of the subarray actual sensors.

Key words: 2-D virtual ESPRIT algorithm, spatial smoothing, virtual array transformation, DOA matrix method

摘要： 在二维虚拟ESPRIT算法的基础上，提出了一种高性能的改进算法。改进方法依照子阵数据直接空间平滑的思想对子阵数据进行预处理，用虚拟阵列变换技术实施阵列变换，采用波达方向矩阵法的思路计算二维角度参数。仿真结果表明，相比于空域平滑的DOA矩阵法和空域平滑的二维虚拟ESPRIT算法，该方法在低信噪比情况下对相干信号源具有更好的估计性能，对独立信号源，能够估计出超过子阵阵元数的信号。

关键词: 二维虚拟ESPRIT算法, 空间平滑, 虚拟阵列变换, 波达方向矩阵法