对传感器阵列信号波达方向(DOA,direction of arrival)估计算法研究多集中在均匀线阵、圆阵等规则阵型阵列,但在水下传感网络等应用环境中阵元分布具有一定随机性,阵列呈现不规则特点.论文利用虚拟阵元思想,通过对不规则阵列的虚拟内插,在逻辑上构建多个具有相同配置参数的子阵列;对多个结构相同的子阵,利用空间平滑技术进行相关目标方位估计;为消除内插引入的非高斯有色噪声影响,引入信号高阶累积量,用累积量矩阵代替协方差矩阵,保证算法精确高效.通过计算机仿真,首先对比了不同信噪比和快拍数条件下提出的算法与传统算法性能,其次研究了构建子阵时角度间隔的取值对估计结果的影响,最后定量分析了特定信噪比和快拍数下新算法DOA估计的平均偏差和标准差,结果显示,新算法能有效提高估计精度.
Irregular Array Interpolation Transformation Direction of Arrival Estimation Based on Higher-order Cumulants
Research on sensor array direction of arrival(DOA)estimation mainly focus on regular formation arrays such as uniform linear array and circular array.But in the actual application environment,such as underwater sensor networks,array elements are distributed randomly.This paper uses the thought of virtual array to interpolate the irregular array virtually and constructs sub-arrays with same configuration parameters logically.For the sub-arrays with same structure,DOAs of correlation signals are estimated with spatial smoothing technique.To eliminate the effects of non-Gaussian colored noise caused by interpolation,higher-order cumulants is introduced and signal covariance matrix is instead by cumulants matrix to ensure algorithm accurate and efficiency.With computer simulations,first we compare the performance of the proposed algorithm and traditional algorithm in different SNRs and snapshots.Then we study the impact of angle interval values on estimation results.At last,we analyze the DOA estimation mean deviation and standard deviation quantitatively in specific SNRs and snapshots to determine the applicable environment and occasion of proposed algorithm.The results show that the new algorithm can improve the estimation accuracy effectively.
NETL
NSTL
万方数据
