基于双谱三维向量的雷达辐射源个体识别

Radar Specific Emitter Identification Based on Bispectrum Three-dimensional Vector

摘要:

基于相位噪声造成的辐射源个体差异，提出了一种新的雷达辐射源个体识别算法。该算法以由双谱熵、双谱能量熵以及双谱主成分均值所组成的三维向量为特征向量，以K-均值（K-means）算法为分类器，完成了雷达辐射源个体识别。相比于现在比较主流的使用深度学习模型作为分类器的辐射源个体识别算法，该算法无需繁琐的模型训练步骤也能取得较好的识别效果。仿真结果表明，在信噪比为0 dB时，该算法依然能确保80%的识别正确率，算法性能优越。

Abstract:

On the basis of the differences of emitters caused by phase noise, a new radar specific emitter identification algorithm was proposed. The algorithm used a three-dimensional vector composed of bispectrum entropy, bispectrum energy entropy and bispectrum principal component mean as feature vector, and completed radar specific emitter identification with K-means algorithm as classifier. Compared to the popular emitter individual identification algorithms that use deep learning models as classifiers, the proposed algorithm can achieve good identification results without the cumbersome model training steps. Simulation results show that when the signal-to-noise ratio is 0 dB, the algorithm can still ensure 80% identification accuracy, demonstrating superior performance.