Author(s)

Zijian Wang

Corresponding Author

Zijian Wang

Abstract

This dissertation mainly studies the recognition of modulation modes of digital communication signals, and studies the modulation recognition algorithms from the perspectives of feature extraction and classifier design. The modulation recognition algorithm based on instantaneous characteristics is studied. The algorithm selects the instantaneous characteristics of seven communication signals, including the peak amplitude spectrum and the absolute amplitude standard deviation, as the classification features. Then, the decision criterion of the decision tree classifier is set according to the classification performance of each feature parameter, and the decision threshold is selected in combination with the simulation results to complete the decision. The construction of the tree classifier. The performance of the algorithm is analyzed through simulation experiments. A modulation recognition algorithm based on the joint features of high-order cumulant and entropy is proposed. In this algorithm, in order to improve the noise resistance and stability of the algorithm, high-order cumulant features are extracted to identify digital communication signals, and entropy features are extracted to classify signals that cannot be identified by high-order cumulant features. According to the classification performance of each characteristic parameter, the decision criterion and threshold value are selected to construct a decision tree classifier, so as to realize the recognition of the signal modulation mode. The performance of the algorithm is analyzed through simulation experiments, and the performance is compared with the modulation recognition algorithm based on the instantaneous characteristics of the signal.

Keywords

Communication signal, Recognition algorithm, Model structure