Cooperative modulation identification method based on multi-class characteristic parameters and evidence theory

Abstract

The invention provides a cooperative modulation identification method based on multi-class characteristic parameters and an evidence theory. The method is used for realizing correct identification on 13 modulation modes of 2ASK, 4ASK, 8ASK, 2PSK, QPSK, 8PSK, 2FSK, 4FSK, 8FSK, 16QAM, 32QAM, 64QAM and OFDM13. The cooperative modulation identification method comprises the following steps that: cooperative nodes extract different classes of characteristic parameters respectively based on characteristics of instantaneous information, characteristics of the combination of detail coefficients of wavelet decomposition and high-order cumulants, and characteristics of a signal cepstral coefficient, wherein the characteristic parameters can represent a target signal from different angles; each node inputs the characteristic parameters to a trained BP (back propagation) neural network for testing; and then the output of the neural network is directly used as an evidence and is conveyed to a fusion centre to be fused, and a fusion rule adopts a D-S (Dempster-Shafer) evidence theory combined rule.

Description

technical field [0001] The invention relates to a cooperative modulation identification method based on multi-category characteristic parameters and evidence theory, and belongs to the technical field of communication. Background technique [0002] Modulation recognition is an intermediate step between signal detection and information demodulation, and has a wide range of applications in wireless communication. The modulation recognition technology of communication signals can be roughly divided into decision theory recognition and statistical pattern recognition. The former uses the viewpoint of probability and compound hypothesis testing to obtain the judgment criterion of the classifier. The main difficulty of this method is how to correctly determine the threshold, and the computational complexity is also very large. The other is pattern recognition, which forms a feature vector by extracting the characteristic parameters of the signal, and uses a pattern recognition ...

AI Technical Summary

Problems solved by technology

This method aims at the problem that in the existing collaborative identification methods, the nodes that cooperate with each other are limited to extracting the same type of characteristic parameters, but the types of signals represented by a single type of characteristic parameters are limited, resulting in very limited types of signals that can be identified by the system. Combining different types of feature parameters and selecting appropriate fusion rules to increase the types of identifiable signals of the system without reducing the average recognition rate of the system

Images