一种基于小生境遗传算法的星座识别方法

Novel Constellation Recognition Method for PSK/QAM Modulations Based on Niche Genetic Algorithms

提出了一种新的PSK/QAM星座识别方法,将星座聚类问题转化为星座二维密度函数的数值优化问题,采用小生境遗传算法对该多峰值函数数值矩阵进行有效搜索,获取星座中心,最后结合聚类有效性评价和星座模板匹配,实现了信号星座类型的识别。与基于模糊C-均值的星座聚类方法相比,新方法与初始聚类中心和样本输入次序无关,缩减了星座识别过程中对聚类结果的评价,减小了运算量。

A new constellation identification method was proposed, in which the clustering problem was converted to numeric optimization of constellation＇s 2-D probability-density-function, and then was resolved by a Niche based genetic algorithm, thus signal＇s constellation was recognized with the help of clustering-validity evaluating and pattern matching. Comparing with traditional clustering algorithms based on Fuzzy C-Means, this presented method is irrelative with initial centers and samples＇ input order, and costs less in evaluating clusters.