Joint LeNet and Knowledge Graph Approach to Link Building Behavior Recognition

In the electromagnetic countermeasure environment, communication behavior recognition is an important part of the mining and utilization of electromagnetic space signals. In order to solve the problem of the difficulty of link establishment behavior recognition and detection in non-cooperative wireless networks and the weak feature representation ability of traditional methods, this paper proposes a link establishment behavior recognition method combining LeNet and knowledge graph, which establishes a feature database based on data preprocessing and uses knowledge graph to visualize the features. In addition, the method can also have a good recognition effect on IQ signals under the condition of low signal-to-noise ratio, and the average accuracy rate of the signal-to-noise ratio is more than 90% when the signal-to-noise ratio is -20 dB on the basis of the measured data, which can effectively identify the link establishment behavior.