Recognition Of Communication Signal Modulation

Modulation method is one of important characteristic of communication signal. It is an important research area of software radio, communications reconnaissance, electronic warfare, radio monitoring etc. It have to determine the modulation method and modulation parameters in the complex and noise jamming signals environment. So as to provide a basis for further analysis and processing. Modulation method is an important difference between different communication systems. The aim of recognition is to identify the modulation method through analysis and processing the received signal absence of prior information. With the development of communication technology, come out a large variety of modulation technology, which give people a higher demand.In this paper, make deeply research on the current principal method of identification. Analyse the problems and give the ideas to improve. The main work can be summarized as follows:Firstly, a brief relevant theory of modulation recognition, then a more detailed description of the common characteristics and signal model. For the problem of signal feature extraction in time/frequency domain, give a way on variable sliding window, theory and simulation results are given. The method have characteristics of simple, high-precision, a few parameters. It reduce the manual setting in the process of identification.Secondly, because the same constellation diagram of QPSK and OQPSK,8PSK and pi/4QPSK signal, the similar time/frequency domain characteristics and the similar HOC, give a classification method based on the phase-hopping statistical characteristics. The method seizes the essential difference between these two sets of signal, extracted the important feature of "number of phase hopping". The simulation show that it is validity by this method.Finally, Give a revised program based on the result achieved primary combined author's work. Calculate the typical HOC then train the RBF network, by this way can recognize signals preliminary. And then recognize further through combined with time domain signal characteristics, frequency domain characteristics ,power spectrum and other characteristics. Achieved good results in a large number of simulation.