摘要
在低信噪比条件下,对6种数字通信信号的自动识剐进行了研究,根据数字通信信号高阶矩和高阶累积量的特性提取了1组特征参数,采用分层结构的BP神经网络进行识别。仿真结果表明:当样本数据足够多,信噪比为4 dB时,正确识别率接达98％。
This paper studies an automatic identification method for six kinds of digital communication signals in the condition of low signa-to-noise ratio. The feature parameters that are higher-order moments and cumulants of digital communication signals are derived from the received signals. Then a hierarchical artificial neural network classifier is used to accomplish the recognition. The simulation results show that the correct identification rate reaches 98% with 4 dB of signal-to-noise ratio and enough data samples.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2006年第5期608-611,共4页
Journal of Nanjing University of Science and Technology
关键词
信号识别
高阶累积量
人工神经网络
signal identification
higher-order moment
artificial neural networks
