ILS通信导航信号调制半监督识别算法仿真  

Simulation of Semi-Supervised Recognition Algorithm for ILS Communication and Navigation Signal Modulation

作　　者：万棣 焦伟东 WAN Di;JIAO Wei-dong(College of Electronic Information and Automation,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学电子信息与自动化学院,天津300300

出　　处：《计算机仿真》2024年第11期96-101,共6页Computer Simulation

基　　金：2022年中国民航大虚拟仿真实验教学建设项目(XF202205)。

摘　　要：ILS通信信号接收识别具有重要的实际意义,但随着空间电磁场日益复杂化与多元化,传统模板匹配算法无法有效识别出畸变的雷达信号。为提升信号识别的稳定度与准确率,基于改进双特征提取融合算法,构建出SCLS-SVM调制信号智能识别模型。首先采用CWD算法调制5类单信号与3类组合信号,并利用图像处理的方法增强特征提取效果;然后利用改进D-CNN联结网络提取调制信号的时频图特征,同时采用优化SA-LSTM算法提取调制信号的时域信息特征;接着将特征融合的数据划分成训练集与测试集,完成信号识别输入建立;最后基于支持向量机分类器,利用十折交叉验证优化的方式,构建SCLS-SVM信号识别模型。消融实验结果表明,SLS、DCN与CVO优化策略的叠加有助于信号识别模型的性能提升;对比实验结果表明,在MSD数据集上,较其它基线识别算法相比,SCLS-SVM算法的P与R评价指标平均增加了6.75%和9.97%,具有最高的稳定性与准确率。综上,SCLS-SVM算法通过改进双特征提取融合算法,提升了调制信号的可辨识度,且SVM算法增加了特征的分类识别效果,在ILS通信信号仿真识别领域具有重要的研究价值。The reception and recognition of ILS communication signals have important practical significance,but with the increasing complexity and diversity of spatial electromagnetic fields,traditional template matching algorithms are unable to effectively identify distorted radar signals.To improve the stability and accuracy of signal recognition,an intelligent recognition model for SCLS-SVM modulated signals is constructed based on an improved dual feature ex-traction fusion algorithm.Firstly,the CWD algorithm was used to modulate five kinds of single signals and three kinds of combined signals,and the image processing method was used to enhance the feature extraction effect;Then the time-frequency feature of the modulated signal was extracted by the improved D-CNN,and the time-domain information feature of the modulated signal was extracted by the optimized SA-LSTM algorithm;Next the data of feature fusion was divided into training set and test set,and the input of signal recognition was established.Finally,based on the support vector machine classifier,the SCLS-SVM signal recognition model was constructed by using the ten-fold cross-validation optimization method.The ablation experiment results show that the superposition of SLS,DCN and CVO optimization strategies is helpful to improve the performance of the signal recognition model;the comparative ex-periment results show that on the MSD data set,compared with other baseline recognition algorithms,the P and R e-valuation indexes of SCLS-SVM algorithm are increased by 6.75%and 9.97%on average,and it has the highest sta-bility and accuracy.To sum up,the SCLS-SVM algorithm improves the identifiability of modulated signals by impro-ving the fusion algorithm of double feature extraction,and the SVM algorithm increases the classification and recognition effect of features,which has important research value in the field of ILS communication signal simulation and recognition.

关 键 词：通信信号调制 特征提取 支持向量机 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]