振速轴向不一致下矢量传感器阵列方位估计方法

Direction of arrival estimation approach via a vector sensor array under velocity axial inconsistency

针对振速轴向不一致时矢量传感器阵列的方位估计精度不高问题,提出了一种两步加权交替迭代自适应方法(two-step weighted alternating iterative approach,TWAIA)。在矢量传感器阵列稀疏信号模型中引入轴向角度偏差参数,并把重构的干扰加噪声协方差矩阵作为加权项,基于加权协方差矩阵拟合和加权最小二乘分别构建了关于稀疏信号功率和轴向偏差矩阵的代价函数。首先固定轴向偏差矩阵,采用正则化加权稀疏协方差矩阵拟合方法估计稀疏信号功率;然后固定稀疏信号功率,采用正则化加权最小二乘估计轴向偏差矩阵,并根据轴向偏差在矩阵中的分布特性,重构期望的轴向偏差矩阵,以此交替的方式迭代更新稀疏信号功率和轴向偏差矩阵,直到被估计的稀疏信号功率相较于前一次的迭代值不再变化为止。最终通过对估计的稀疏信号功率谱峰搜索即可实现声源的波达方向估计。仿真结果表明,相较于现有方位估计方法,提出的TWAIA提高了振速轴向不一致时矢量传感器阵列的方位估计精度。

To improve the direction of arrival accuracy via the vector sensor array in the presence of velocity axial inconsistency,a two-step weighted alternating iterative approach(TWAIA)was proposed to estimate the DOA.First,the axial angle bias parameter was introduced into the signal model of the vector sensor array,and the reconstructed interference plus noise covariance matrix was used as the weighting item.Then,based on the weighted covariance matrix fitting and weighted least squares,the cost functions with respect to the sparse signal power and the axial angle bias matrix were reconstructed.In the first step,the axial angle bias matrix was fixed,the regularized weighted sparse covariance matrix fitting method was used to estimate the sparse signal power.In the second step,the sparse signal power was fixed,the regularized weighted least squares was employed to estimate the axial angle bias matrix.According to the bias distribution characteristics in the axial angle bias matrix,the desired axial angle bias matrix was reconstructed,and then both the sparse signal power and the axial angle bias matrix were iteratively updated in an alternating way until the estimated sparse signal power was no longer changed compared with the result of the previous iterative estimation.Finally,the direction of arrival(DOA)was obtained by searching the estimated sparse signal power spectrum peak.Simulation results show that,compared with existing methods,the proposed TWAIA improves the DOA estimation accuracy of the vector sensor array under the condition of velocity axial inconsistency.