摘要： 提出了一种基于星座聚类的MQAM调制识别新方法，运用一种改进的基于遗传算法的自适应聚类算法对MQAM星座进行重构和识别。该自适应聚类算法利用遗传算法的高效全局搜索特性，克服了模糊C-均值算法对初始聚类中心和样本输入次序敏感等不足，结合聚类有效性分析实现了聚类中心数目的自适应调整。仿真结果表明，基于该聚类算法的MQAM信号调制阶数识别方法是有效的。

关键词: 遗传算法, 聚类, 星座, 调制识别

Abstract: A novel modulation recognition method for MQAM signals is proposed. It employs a modified genetic algorithms (GA) based clustering method to rebuild and recognize constellations. The presented clustering algorithm not only overcomes the sensitivity to initial centers as fuzzy C-means with the effective global searching ability of GA, but also determines the number of clusters adaptively via validity analysis. Experiments show that this MQAM recognition method is effective.

Key words: genetic algorithms, clustering, constellation, modulation recognition

中图分类号: 

• TP312