A new method of automatic digital modulation recognition based on ART2A-E neural network has been proposed after analyzing the traditional structure of ART2/2A neural network, and the ability of the proposed classifier are evaluated with six different modulation types such as 2PSK,4PSK,OQPSK,GMSK,π/4OQPSK and 4ASK.
In view of signal characteristic about spectrum correlation and based on the spectrum correlation theory, this article has studied a modulation recognition algorithm which may distinguish AM, CW, FM, DSB, VSB, LSB, USB, 2PSK, 4PSK, 2ASK, 4ASK, 2FSK and 4FSK modulated signals.
This article mainly has an analysis in three aspects: one is based upon signal' s own characteristic parameters, the other relies on the parameters gotten by wavelet analysis and the final is the application of artificial neural network in automatic modulation recognition.
Automatic modulation recognition is an important research area for the receiver design in the non-cooperative communication system. A digital modulation recognition algorithm is proposed in this paper for the engineering application.