Study Of Digital Modulation Signal Recognition Algorithm

Modulation Classification of Digital modulation signals is commonplace in both commercial and military systems. This dissertation studies on the recognition of digital modulation signals. The main work of this dissertation is as follows:1.This thesis compares four different digital modulation signals in both time and frequency domains. Besides, the estimation methods of carrier frequency, the ratio of signal to noise (SNR) and data period have been also covered. An algorithm based on maximum peak of power spectrum of signal's delay product is adopted to estimate data period, which is verified by simulations.2.The recognition method based on statistical feature parameters for different signals has been analyzed in detailed. And this technique is simple and easy to operate. What's more, another approach named spectral correlation of modulated signals is studied and compared, which is proven to be time-consuming and low-efficient. Finally, as to six different digital modulated signals, five feature parameters are constructed for the signal recognition using the high order cumulants classification approach.3.Chapter 4 theoretically analyzes the influences of effect of roll-off factor on MQAM signals, and conclusions have been made for their relations by virtue of computer simulations. Moreover, the recognition of MQAM signals based on Hilbert transform have also been investigated, which is, in contrast, largely dependent on the SNR properties.4.An improved algorithm for MQAM signal recognition is introduced: the method firstly uses two parameters (R and M) to recognize 4QAM and 16QAM signals; then based on Hilbert transform, the left 32QAM and 64QAM signals are also extracted. Experiments indicate that this new method accelerates the signal processing speed, and is still accurate and effective under low SNR. Therefore, the proposed approach could be applied in the fast detections and recognitions of communication systems.