Classifier of modulation recognition based on modified self-organizing feature map neural network

In order to improve recognition probability and decrease recognition time of the classifier of modulation recognition, and to make it adaptive, the self-organizing feature map neural network (SOM) was used as the classifier of modulation recognition, applying its property of self-organizing and adaptation to adapt automatically variety of the signal to noise ratio. Learning rule and competitive transferring function were modified in order to have two victorious output neurons. The number of output neurons could be decreased and the convergent rate of neural network could be improved thought those ameliorations, and modulation type could be recognized in less time. The simulation results proved that recognition probability of modified SOM is higher than other neural networks, and modified classifier is implemented easily in practical engineering because of its simple structure.