Modulation Recognition of MQAM Signals Based on Semi Supervised Clustering Theory

Abstract

In the modulation recognition of MQAM signals cluster points of traditional clustering algorithm were not accurate, iterations of the algorithm are multiple and the curve of square error function was not smooth. To solve these problems, this paper presents a theory of modulation recognition method for reconstruction of MQAM signal constellation diagram based on semi supervised clustering, using labeled samples to guide the membership degree and the clustering center update. Analysis the receiving constellation and extracting the characteristic parameters R of constellation compared with parameter Rs of standard constellation, to realize modulation recognition of the different order of MQAM signal. The results show that the method to identify the MQAM signal at rate of 90. Convergence of iterative process is reduced from 40 to 8 times. The algorithm has low computation complexity and the square error function curve is smooth.