Design And Realization Of Automatic Modulation Recognition Of Digital Communication Signals

The automatic modulation recognition of communication signals is one of the important communications issues. As early as 60 years in the 20th century, people began to explore this area. With the development of modern communications technology and putting forward the idea of Software Radio, the automatic modulation recognition of communication signals were gradually concerned by more and more people. Since the nineties of the twentieth century many articles have been published by the researchers around the world. They made a lot of good algorithm and promoted the rapid development in this area. Based on previous studies, this paper proposed a n automatic classification algorithm for digita communication signals which can effectively distinguish the ASK, FSK and PSK signals with high accuracy and calculation speed.Power spectrum can reflect many features of modulated signals. Despite the change in carrier frequency, symbol rate and SNR's will affect the power spectrum, the shape's characteristics of the same modulation signal's power spectrum remained. In this paper, the algorithm is mainly based on characteristics of the signal's power spectrum. With proposed five parameters, the algorithm can successfully identify CW, MASK, MFSK, BPSK and QPSK signal.Because of not using characteristics sensitive to symbol rate and noise such as Instantaneous phase, Instantaneous frequency, the algorithm can performance well in a wide range of symbol rate and noise. Compared with methods of cycle spectral, the wavelet transform and the high-order signal processing, the algorithm of this paper has a faster response without complex mathematical operations. The five parameters were provided as training samples to the SVM classifier after the simulation of their distribution. Simulation results show that, in the same classifier, the overall recognition accuracy of the modulation signals was higher than 98%.Finally, a demonstration system of automatic modulation recognition of digital communication signals was shown in this paper. The system includes the following modules: AD acquisition, system control, signal processing and display. Through putting the algorithm into the DSP and training the SVM with actual signal, the demonstration system can successfully identify the modulation signals with center frequency of 21.4MHz. Therefore, the system further validates the effectiveness and the realizable of the algorithm on the basis of simulation.