针对分布式阵列的波达方向估计,提出了一种基于多尺度旋转不变(MI)MUSIC的波达方向估计方法。首先对分布式阵列进行子阵划分,构造具有多尺度旋转不变性的子阵;然后利用酉ESPRIT算法得到精度低但无模糊的粗估计,采用MI鄄MUSIC算法得到一组精度高但包含模糊的精估计;最后以粗估计为参考解精估计的模糊,从而得到高精度无模糊的波达方向估计。所提方法具有分布式阵列及MI-MUSIC的优点,对子阵内部阵元位置误差不敏感,且精度较双尺度ESPRIT高。计算机仿真结果验证了该方法的有效性,也验证了分布式阵列DOA估计中存在基线模糊门限与信噪比门限。

An algorithm based on MI-MUSIC is proposed for DOA estimation of distributed arrays. Firstly, conduct subarray division to distributed arrays to construct multiple invariance subarrays. Secondly, the Unitary-ESPRIT algorithm is used to obtain low-accuracy but unambiguous coarse estimates, and the MI-MUSIC is used to obtain fine estimates of high-precision, yet containing ambiguity. Finally, coarse estimates are used as a reference to resolve the ambiguity of fine estimate, thereby high-precision and unambiguous DOA estimations can be obtained. The proposed method has the advantages of distributed arrays and MI-MUSIC. It is not sensitive to the element misalignment and has a higher angular accuracy than that of dual-size ESPRIT. Simulation results demonstrate high accuracy of DOA estimation of the proposed method, and that baseline threshold and SNR threshold are existed in the distributed arrays.