Modulation Classification Based On Fuzzy System

Modulation identification of communication signals is developed rapidly in recent years, and it is a critical part for signal processing. The objective of automatic modulation identification is to determine the modulation type and estimate the modulation parameters without any priori knowledge. For vague information or corrupted signals, fuzzy logic has the advantages for its simplicity and flexibility.In this paper, the basic theory of the fuzzy logic system is introduced, including the concept of fuzzy sets, membership function, and the rules of fuzzy inference system. The structure of Mamdani is discussed and the implementation of the structure by Matlab is particularly analyzed. The automatic modulation identification of communication signals based on fuzzy logic classifier is realized. It is shown by the simulation results that the percentage of classification identification is no less than 79.8% when SNR is 5dB and the percentage of classification identification is 100%. As the membership function is influenced severely by the experience of the operator, a new algorithm is proposed based on the FCM (the cluster of fuzzy C mean). The algorithm gets the center of cluster firstly and then the modulation type of the signals can be determined according to the distance between the center and the signal. It is shown by the simulation results that the percentage of correct identification is no less than 100% when SNR is 5 dB. In conclusion, the fuzzy system can be used to the automatic modulation identification.