Modulation Recognition On Digital Conmunication Signals

Modulation recognition on digital communication modulation signals is an important issue of signal processing study, widely used in military and civilian fields. This paper studies the parameter estimation and the modulation recognition using decision theory on the digital modulation signal. Modulation recognition algorithms of the MQAM signals is analyzed, By improving the method of characteristic parameters extraction, we achieve the recognition of more signal types.This issue introduces the basic digital modulation characteristics in the time domain and frequency domain and achieves the modulation simulation of common digital communication signals; Hilbert transform is used to extract the instantaneous characteristics of digital signals and the carrier frequency estimation and SNR estimation are also discussed. The symbol rate estimation method based on wavelet transform is introduced according to the multi-resolution characteristics of wavelet transform. After two continuous wavelet transform and a Fourier transform, the symbol rate of the digital communication modulation signals is extracted with the purpose of modulation recognition.According to the previous instantaneous characteristics, we select and analyze the five parameters with good discrimination in detail, and using the threshold established we use the decision theory to identify the six common digital signals including 2ASK, 4ASK, 2FSK, 4FSK, 2PSK and 4PSK. We introduce MQAM signal characteristics theoretically and discuss the roll-off characteristics of MQAM signals, Then MQAM signal recognition method based on Hilbert transform is studied, with the performance evaluation at different SNR behind. This methods is applicable to identify MQAM signals in the category, but it has some limitations. Next, we improve the extraction of characteristic parameters, and achieve the eight digital communication signals including 8PSK, 16QAM using the application of new six parameters. This method reduces computation, easy to program, and improve the recognition accuracy in the applicable SNR range.