基于自适应滤波算法的信号处理研究

Research on Signal Processing Based on Adaptive Filtering Algorithm

通过接收到的回波信号,波达方向定位技术(DOA)可以获取目标的距离信息和方位信息。传统的基于子空间分解的DOA估计方法需要对空间协方差矩阵进行特征值分解(EVD),方法计算量大并且难以实现实时应用;基于固定步长最小均方差(FSS-LMS)算法的DOA估计方法可以有效降低计算量,为此,文章提出了一种基于可变步长最小均方差(VSS-LMS)算法的DOA估计方法。利用估计出的误差信号来更新步长,通过阵列模式的倒数得到空间频谱,其峰值表示信号的DOA估计。仿真结果表明,与FSS-LMS算法和矩阵特征空间分解(MUSIC)算法相比,VSS-LMS算法的DOA估计方法能以较低的计算复杂度实现较好的性能。

Direction of arrival(DOA)technology can obtain the range information and azimuth information of the target through the received echo signal.The traditional DOA estimation method based on subspace decomposition needs to do eigenvalue decomposition(EVD)of spatial covariance matrix,which is computationally expensive and difficult to realize in real-time application.The DOA estimation method based on fixed step least mean square(FSS-LMS)algorithm can effectively reduce the computational complexity.This paper presents a DOA estimation method based on variable step size least mean square algorithm(VSS-LMS).The step size is updated by the estimated error signal,the spatial spectrum is obtained by the reciprocal of array mode,and the peak value represents the DOA estimation of the signal.Simulation results show that,compared with FSS-LMS algorithm and music algorithm,the DOA estimation method of VSS-LMS algorithm can achieve better performance with lower computational complexity.