未知噪声环境方位-俯仰-频率-时延等四维参数联合估计方法

提出一种基于空时扩展虚拟传感器阵列的未知噪声环境下方位、俯仰、多普勒频移和相对时延四维参数联合估计新算法。未知噪声背景下的UN-MUSIC、UN-CLE等算法不易于完成多参数联合估计任务,而UN-ESPRIT算法又需要信号具有空、时域子空间旋转不变特性,对阵列结构要求严格。文中通过对原始数据的时延补偿等处理,利用虚拟数据阵的时移旋转不变特性,经由构建空时扩展波达方向矩阵同时获得诸参数的联合估计。虚拟传感器阵列的理论孔径为物理阵列的数倍,具有在低信噪比、未知分布噪声环境下更强的适应能力;且算法对于物理阵列阵元分布无特殊要求,具有很好的理论和实用价值。理论分析以及计算机仿真都证明了算法的有效性。

Joint Azimuth-Elevation-Frequency-Delay Estimation Based on Virtual Arrays in Unknown Noise Environment

Based on a pair of tempo-spatial expanded virtual sensor arrays, a novel algorithm for 2-dimentional Direction of Arrive (DOA) angle, Doppler frequency and relative time-delay estimation is proposed in unknown noise field.Those sub-space methods for unknown noise environment such as UN-MUSIC and UN-CLE are not suitable for multi-parameter jointly estimation, and the structure of arrays would be too complicated when use UN-ESPRIT method.In this paper,by the time-delay compensating of original data,a tempo-spatial expanded DOA matrix is built by the temporal rotational invariance of virtual data matrixes, and from this matrix the estimations of 2D-DOA, frequency and time-delay can be completed simultaneously.The large aperture of virtual array make it more suitable for parameters estimation even in heavy unknown correlative noise fields. This algorithm can also be used to any plane sensor array such as uniform Circle arrays and uniform L-shape arrays. Simulations are presented to show the capabilities and performance of this method.