基于数据协方差矩阵重构的MIMO声纳DOA估计

实际多输入多输出(MIMO)声纳系统由于环境或人为因素,可能出现部分阵元失效,从而导致阵列自由度减少、方位估计精度下降。本文提出了一种数据协方差矩阵重构方法,该方法基于差分阵列性质,利用正常工作阵元的协方差矩阵元素来恢复失效阵元的矩阵元素,获得满秩的数据协方差矩阵,从而恢复到全阵元MIMO声纳的阵列自由度。与已有方法相比,降低了计算复杂度。仿真及海试实验数据处理结果表明,本文所提的数据协方差矩阵重构方法能够恢复因部分阵元失效而丢失的阵列自由度,应用于方位估计中,所能分辨的最大目标数与全阵元相同。

DOA estimation in MIMO sonar based on sample covariance matrix reconstruction

Due to the complexity underwater environment and human factors, some sensors may be broken in the Multiple-input multiple-output (MIMO) sonar. It will decrease the degree of freedom (DOF) and direction-of-arrival (DOA) estimation precision. In this paper, a covariance matrix reconstruction method is proposed to overcome the problem of missing data from broken sensors in MIMO sonar. Firstly, based on the property of difference co-array, the data elements corresponding to the broken sensors in the covariance matrix are constructed from the intact sensors. Then a new covariance matrix of MIMO sonar with full-rank is obtained. That is, DOF is restored and equal to the DOF of intact MIMO sonar, as well as the performance of DOA estimation. Comparing with other reconstruction methods, the proposed method reduces the mathematical complexity. Numerical simulations and experimental results show that the effectiveness of the proposed method. In DOA application, it has the same maximum resolvable number of sources as the intact MIMO sonar.