基于次分量分析的DOA跟踪算法  

DOA tracking based on minor component analysis approach

作　　者：刘成城[1] 李冬海[1] 赵拥军[1] 

机构地区：[1]解放军信息工程大学信息工程学院,郑州450002

出　　处：《计算机应用研究》2010年第7期2492-2493,2519,共3页Application Research of Computers

基　　金：国家"863"计划资助项目(2009AA7034530)

摘　　要：针对信号源入射角时变的情况,分析了基于幂迭代的子空间跟踪算法,提出一种次分量分析方法的子空间跟踪算法。该算法首先利用基于反Hebbian学习的神经网络,抽取多个次分量,操作相对简单、算法稳定、收敛快,且有自组织特性;然后提出一种实时并行处理方法,在抽样结束时数据处理完成;最后采用牛顿法实现运动目标的DOA跟踪。仿真实验证明其收敛快、跟踪性能好。As the direction of signal source varies with time,this paper proposed a tracking method based on minor component analysis approach.To extract the minor component,the algorithm was based on anti-Hebbian learning neural network and contain only relatively simple operations,it was stable,converges and had self-organizing properties.Then proposed a method of real-time parallel processing,at end time of sampling data,processing also could be finished.Finally,using Newton algorithm to track moving target.Computer simulation results demonstrate the effectiveness of the proposed algorithm.

关 键 词：DOA跟踪 幂迭代 次分量 反Hebbian 牛顿法 

分 类 号：TN911[电子电信—通信与信息系统]