Identification Of Digital Modulated Signal

In the modern communications systems, the signal must be modulated on the transmitting terminal before forward, converting into a mode which is suitable for transmitting in the medium. When the signal reaches the receiving end, it will be demodulated and reductive into a baseband signal. Because there are many kinds of modulation in the digital communication system, it is the key to demodulate the signal that identifying the modulation scheme of the signal on the receiving the end.This thesis briefly introduces the importance of accurate identification of signal modulation scheme firstly and describes the current situation for the modulation signal recognition at home and abroad. Then it expounds the basic principles of modulation, focuses on the modulation principle of multi-band digital modulation signal, including the mathematical model of the signal, the model of getting and the time-frequency domain characteristics. In this thesis, it uses two methods, genetic algorithms and cyclic spectrum analysis, to distinguish the modulations of the digital signal together.With the two methods it is easy to achieve the amplitude, frequency and phase information signal.Then based on these information we can use the two characteristic parameters of the digital signal including zero central instantaneous amplitude variance and digital signal normalized of cyclic spectrum spectral coefficients to identify the modulation of the digital signal. Through the joint use of genetic algorithms and cyclic spectrum analysis, to overcome the lack of past studies in which it can only identify a particular modulation scheme.The entire identification process consists of two parts, class identification and inner-class identification. With (4ASK,6ASK,7ASK,8ASK;4FSK,6FSK,7FSK,8FSK;4PSK,6PSK,7PSK,8PSK) a total of12the kinds of signals, I make many experimental simulation in the different signal-to-noise ratio. The identification method has a clear structure, high recognition accuracy, and also has better recognition results in less than5dB signal-to-noise ratio environment.The papers is about a preliminary theoretical studies on multi-band digital modulation signal recognition. Many works should be done in order to use the identification methods in practice.