摘要
通信信号自动分类是一模式识别问题，通常用数字信号处理和模式分类的方法来求解．文中提出了将ＲＢＦ（ＲａｄｉａｌＢａｓｉｓＦｕｎｃｔｉｏｎ）网络方法应用于通信信号自动识别的具体方法．构造了运用ＲＢＦ网络的信号分类的神经网络结构．通过模拟实验表明，由于采用了将信号特征矢量降维的方法，该网络不仅能够很好地完成信号分类，而且具有比传统方法训练速度快、占用存贮空间少、容错性强和易子硬件实现等特点．
Automatic classification of communication signals can be considered as a pattern recognitionproblem. It can be solved by davital signal processing and pattern recognition. This Paper presents a method to solve this problem by using RBF networks. A structure of neural networks isdesigned by means of the RBF network for recognizing communication signals. Simulation experboantat results show that, owing to using feature Valors to lower dimensions, the new networkscan not only accomplish the signal recognition task, but also give a better result than the existingmethods in terms of training speed, memory robustness and realization by using hardware.
出处
《西安电子科技大学学报》
EI
CAS
CSCD
北大核心
1996年第1期112-120,共9页
Journal of Xidian University
