摘要
本文致力于基于神经网络的通信信号调制类型识别器设计研究。论文提出了一种改进的BP神经网络分类器,它采用7个特征参数,可以对CW、2FSK、4FSK、8FSK、2PSK、4PSK、8PSK、8QAM、16QAM、4ASK、8ASK共11种调制类型实现正确分类识别。论文讨论了方案设计,给出了仿真试验结果,并将其与其他神经网络分类器进行了性能比较。
This paper is focusing on a study on the neural network based classifier design of modulation types for communication signals. An improved BP classifier based on artificial neural networks （ANN） is proposed which, by the use of 7 characteristic parameters, could make correct identification among 11 modulation types, i.e. CW,2FSK,4FSK,8FSK,2PSK,4PSK,8PSK, 8QAM, 16QAM ,4ASK and 8ASK. The design procedure is discussed and simulation results are presented. Finally a performance comparison with other neural network based classifiers is also provided.
出处
《微计算机信息》
北大核心
2005年第11S期102-104,共3页
Control & Automation
基金
总装备部探索研究项目基金资助(编号不公开)
关键词
调制类型识别
特征参数
分层结构组合分类器
神经网络
modulation type identification, feature parameter, hierarchical architecture combined classifier, artificial neural networks
