一种基于连续蚁群算法的快速最大似然DOA估计  

Fast Estimation of Maximum Likelihood DOA Based on Continuous Ant Colony Algorithm

作　　者：王辉辉[1] 陈玉凤[2] 

机构地区：[1]西安电子工程研究所,西安710100 [2]西北工业大学航海学院,西安710072

出　　处：《火控雷达技术》2014年第2期1-4,67,共5页Fire Control Radar Technology

基　　金：国家自然科学基金(60972152);国家重点实验室基金(9140C2304080607)

摘　　要：针对传统最大似然估计计算量大的问题,将连续空间蚁群算法与最大似然估计算法相结合,在ACOML算法的基础上,提出了一种用混沌序列初始化状态空间的改进蚁群算法MACOML(Muddleheaded ACO)。该方法使用混沌映射产生的初始状态空间来代替ACOML算法中的随机序列产生的初始状态空间,增加了初始解的遍历性,同时在寻优过程中增加了局部搜索。仿真结果表明:MACOML能保持最大似然估计方法的高分辨性能,而计算复杂度只是最大似然方法的1/20。Aiming at problem of great amount of computation for maximum likelihood estimation, an improvement on ant colony algorithm（ Muddleheaded ACO） by using chaotic sequence to initialize state space is presented on basis of ACOML algorihtm with combination of continuous space ant colony algorithm with maximum likelihood estimation algorithm. The method uses initial state space generated by chaotic map to substitute the initial state space generated by random sequence in ACOML algorithm, and ergodicity of initial solution is added, and local search is added during looking for optimization. The simulation results indicate that using MACOML can retain high resolution capability of maximum likelihood estimating method, and the computing complexity is only 1/20 of that using maximum likelihood method.

关 键 词：最大似然估计 蚁群算法 混沌映射 计算复杂度 

分 类 号：TN911.7[电子电信—通信与信息系统] TP301.6[电子电信—信息与通信工程]