基于两种神经网络方法分析的DOA估计算法比较

许多高分辨率波达方向估计算法如MUSIC和ESPRIT估计都是以子空间概念为基础并且需要输出相关矩阵的特征值分解,由于数量估计的特征值分解计算,因此提出的PCA和MCA是分别基于信号子空间和噪声子空间的估计算法,算法稳定、收敛,且有自组织特性。仿真实验表明两种神经网络DOA估计算法具有不同的性能。

DOA Estimation Algorithm Comparison Based on the Two Neural Network

Many high-resolution DOA estimation algorithm such as ESPRIT and MUSIC are based on the concept of space and need to output the correlation matrix eigenvalue decomposition,the estimated number of eigenvalue decomposition,the PCA and MCA respectively based on signal and noise subspace estimation algorithm stability,convergence,and there is self-organizing characteristics.Simulation results show that both DOA estimation algorithm for neural networks with different performance.