一种基于集成学习的干扰频点检测方法
A Detection Method of Jamming Frequency Based on Ensemble Learning

针对传统干扰频点检测方法所存在的精度不够高、易受噪声干扰、实时性较差问题，提出了一种基于集成学习的干扰频点检测方法。首先，对仿真数据进行归一化、类别标注等处理并生成3种不同规模的数据集；然后，构建多个基于XGBoost的频点干扰二分类模型；最后，通过并联所构建的二分类模型对通信帧数据进行频点干扰识别以得到所有被干扰的频点。仿真实验表明，基于并联XGBoost模型的干扰频点检测在仿真数据测试集上的精确率达96.8%，平均推理时间小于5 ms，验证了所提方法的高效性。
Aiming at the problems of low precision, easily disturbed by noise, and poor real-time performance of traditional jamming frequency detection methods, a new method of jamming frequency detection based on ensemble learning was proposed. Firstly, the simulation data were processed by normalization, category labeling, etc., and three datasets of different sizes were generated. Then, multiple binary classification models based on XGBoost are constructed for a single frequency. Finally, the binary classification model constructed in parallel is used to identify the frequency jamming of communication frame data to obtain all the jamming frequencies. Simulation experimental results show that the proposed method has an average accuracy of more than 96% and an average inference time of less than 5 ms on the simulation data set, which verifies the high efficiency of the proposed method.