基于广义近似消息传递的快速DOA估计方法

针对稀疏恢复类波达方向(direction of arrival, DOA)估计算法中计算复杂度高的问题, 提出了一种基于广义近似消息传递(generalized approximate message passing, GAMP)方法的稀疏贝叶斯学习算法。该算法在现有双基地无源雷达系统模型基础上, 构建了多快拍下的GAMP信号统计模型, 将高维联合后验概率密度的计算简化为标量运算, 提高了算法的计算效率。对于离网目标, 利用梯度下降方法推导了角度空间网格更新策略, 进一步提高了角度估计的精度。仿真结果表明, 该算法在有限快拍、低信噪比情况下, 估计精度较高, 计算复杂度较低, 适用于实时性要求高的应用场景。

Fast DOA estimation method using generalized approximate message passing

In order to solve the problem of high complexity in direction of arrival (DOA) estimation algorithm of sparse recovery, a method is proposed which employs the generalized approximate message passing (GAMP) based sparse Bayesian learning to estimate DOAs. Based on the existing bi-static passive radar system, this algorithm builds a statistical model of GAMP signals with multiple measurement vectors, simplifies the calculation of high-dimensional joint posterior probability density to a scalar operation, and improves the calculation efficiency of the algorithm. In addition, for off-grid targets, the angle spatial grid updating strategy is derived by the gradient descent method. Simulation results show that the proposed algorithm has higher estimation accuracy and lower computational complexity than other algorithms with a limited number of snapshots and low signal-to-noise ratio.