A Research Of Modulation Recognition Based On The Rough Set And The Neural Network

Automatic modulation recognition is an academic focus in the pattern recognition and the signal processing disciplines, which is widely used in electronic warfare, electromagnetic spectrum management and monitoring as well as signal interference and recognition fields, and also with practical military value and civilian value.In the past few decades, many scholars at home and abroad had a in-depth research about the automatic modulation and recognition, and proposed lots of classic methods. Currently, the use of neural networks to achieve automatic recognition is the mainstream method. However, due to the diversity and continuity of feature parameters, the modulation recognition was largely restricted in practical application. After studying masses of academic reports and papers about modulation recognition at home and abroad, also owing to the fact that the rough set theory possesses powerful data mining capabilities and ANN which is well know for its superior fault tolerance, a method that combines the neural network and rough set is given in this paper, and the main works can be summarized as follows:(1) This paper analyzes the signals's instantaneous amplitude, frequency and phase, and also makes a thorough study in theory and experiment.(2) Self-organizing neural network and the rough set are used in this paper to reduce the feature parameters. The simulation shows that the method in this paper can not only reduce the number of feature parameters, but also remove effectively the redundant information. (3) The identification process is given in this paper which combines the rough set and ANN. The experimental result shows that the method can shrinks the network training time and improves the recognition rate and the stability.