The channel of burst communication has the property of low SNR and high-dynamic changes, so as to degrade the performance of the system, therefore, it is very important for the burst communication system to improve the performance of modulation recognition under low SNR. In this dissertation, modulation recognition based on stochastic resonance is studied, Simulation shows that this method can effectively promote the recognition rate under low SNR.The main contributions of the dissertation are as follows:Firstly, a broad overview of modulation recognition is introduced. In section II, Digital signal modulation technique is summarized, and then the cumulants of the modulated signals are analyzed. In Section III, the theories of classifier of neural network and support vector machine are introduced separately. Afterwards, based on the two classifiers and the cumulants, recognition simulation of digital signals of BPSK,QPSK,2FSK,4FSK is implemented. The system of stochastic resonance is presented in section IV, and finally the modulation recognition simulation of the 4 signals above based on stochastic resonance is carried out. The result of the simulation shows that the recognition rate is 80% at the SNR of -6dB. Compared to the previous method, the performance of this method is improved by 6dB and precise recognition can be achieved under Low SNR. |