Abstract:

By statistics of instantaneous characteristic and analysis of characteristic parameters, an identification method was obtained, which is used to recognize the digital communication signal modulation based on the decision theory and the previous research of Azzouzi. The new method solves the probloms,of small identify types, complex procedures and low recognition rate. The new method has great practical significance. The key of the algorithm is to get the best thresholds and the thresholds are variable in different SNR (Signal to Noise Ratio). The paper compares the probability of the characteristic parameters based on the ML（Maximum Likelihood） to obtain the best judge thresholds. The result indicates that this method can recognize 7 different kinds of signals including noises, and the success rate is higher than 96% when SNR is 10 dB, and the success rate is higher than 92% when the SNR is between 6 dB and 15 dB.

Key words: modulation recognition, analysis of characteristic parameter, judge thresholds, maximum likelihood(ML）