中文摘要：

目的提出一种基于粒子群和减法聚类相结合的算法优化聚类半径的调制识别新方法。方法首先使用循环逼近法从接收信号中估计载波频率，然后进行下变频、小波消噪和均衡得到基带信号，最后使用粒子群和减法聚类相结合的方法搜索使重构星座图最优的减法聚类的聚类半径，以得到的聚类半径为分类特征来实现MQAM调制信号的识别。结果使用该方法，实现了4种不同阶数QAM信号的识别。结论该算法和现有减法聚类算法相比识别率有显著的提高，同时该识别方法对低信噪比有一定的健壮性。

英文摘要：

Aim To propose a novel algorithm for modulation recognition of MQAM signals which combines sub- tractive clustering （SC） and particle swarm optimization （PSO） to extract the discriminating features. Methods Firstly, the iteration approximation method is proposed to estimate the carrier frequency. Then after down conversion, wavelet denosing and equalization are performed to get the baseband signal. Finally, PSO is used to search for the best clustering radius of SC in order to enable reconstructed constellation optimal and subsequently the best clustering radius is utilized as a classification feature. Results The proposed algorithm realizes the identification of four different orders of QAM signals. Conclusion The algorithm p by this paper has higher correct classification rate in modulation classification for MQAM signals. In addition, simulation results show that the modula- tion classification method is robust in the low signal-noise ratio.