Automatic Recognition Of Digital Modulation Types Based On Spectral Analysis

Modulation recognition and parameter estimation is an important field in electronic countermeasures, it also is a rapidly evolving field of signal analysis. This paper studies the spectrum ,power spectrum and spectral correlation of several kinds of digital communication signals. The author presents new algorithms which can classify the modulation types and estimate the modulation parameter of a low SNR signal. Because the spectral correlation function of signal is not sensitive to the stationary noise and interference, So the theory of spectral correlation has some advantage than other theory when at a low SNR. When signals(such as QPSK and OQPSK) have the same power spectrum, it is difficult to distinguish them by common approach. But it is easy to distinguish them by approach of spectral correlation because their spectral correlation functions are different obviously. At the same time when using the spectral correlation, it also has its disadvantage such as its algorithms are too complicate to be real-time processed easily. When some signals have the same spectral correlation functions, such as 2ASK and 4ASK,BPSK and QPSK, it is also difficult to distinguish them by spectral correlation approach. SO this paper uses the theory of spectrum, power spectrum and spectral correlation together.Futhermore, this paper studies the use of spectral correlation approach to estimate the carrier frequency and symbol rate of communication signals. The details of algorithm have been given in this paper. Different modulation types can be distinguished by these new algorithms, such as AM, FM, DSB, SSB, BPSK, QPSK, OQPSK, 8PSK, 2FSK, 4FSK, 2ASK and 4ASK. The results show that the effect of the parameters extracted in this text is very good when they are used to distinguish signals.