摘要
传统的聚类算法用在MQAM信号的调制识别中,算法的迭代次数多,特别对高阶调制信号运算时间长。针对该问题,提出了一种改进的半监督聚类重构星座图的方法,用标记的样本点来指导隶属度和聚类中心的更新,降低了算法的运算复杂度,减少了迭代次数,聚类中心数目准确。通过分析接收端星座图,提取星座图的特征参数R并与标准星座图的参数Rs进行比较,实现了MQAM信号调制方式的识别。仿真结果表明该方法对MQAM信号的识别率在90%之上,且算法的复杂度低,尤其当调制阶数较高、数据长度较长时,能够将运算时间减少为原来的1/3。
Traditional clustering algorithm is used in MQAM signal modulation classification. The number of iterations is more, particularly to the high- er-order modulation signal has high time complexity. To solve this problem, a method of improvement semi-supervised clustering to reconstruct constella- tion diagram is presented, using marked sample points to guide the update of cluster centers and membership. The improved algorithm reduces the time complexity and the nunther of iterations. By analyzing the receiving end constellation diagram, extracting characteristic parameters R and comparing to R of standard constellation diagram, achieving MQAM signals modulation recognition. Experiments show that the recognition rate of MQAM signal is above 90% and the algorithm＇s complexity is low. Especially when the modulation order is higher and the data length is longer, the operation time is reduced to the original＇ s 1/3.
出处
《电视技术》
北大核心
2014年第11期112-115,共4页
Video Engineering
基金
国家自然科学基金资助项目(U1204604
61172086)
中国博士后基金资助项目(2012M511587)
河南省博士后基金资助项目(2011829)
河南省青年骨干教师资助项目(2013GGJS-002)
关键词
调制识别
半监督聚类
MQAM信号
modulation identification
semi-supervised clustering
MQAM signal
