集成学习在高误码率下AOS协议识别中的应用研究  

Application of Ensemble Learning in AOS Protocol Recognition with High Bit Error Rate

作　　者：朱明 王春梅 姚秀娟 李雪 ZHU Ming;WANG Chun-mei;YAO Xiu-juan;LI Xue(Nation Space Science Center,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院国家空间科学中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《宇航计测技术》2020年第3期80-87,共8页Journal of Astronautic Metrology and Measurement

基　　金：中国科学院创新基金资助项目(Y42613A32S)。

摘　　要：针对天基网络中高误码率的传输特点,为保证各个异构网络之间数据能够高效可靠的传输,采用空间链路层高级在轨系统协议设计一种基于集成学习的识别方法。该方法采用集成学习模型,学习AOS协议数据,构建基于集成学习的AOS协议识别模型,实现AOS协议的准确识别,并在高误码率情况下进行实验验证。实验结果表明,集成学习模型在AOS协议识别方面具有较好的识别效果,识别运行效率有显著提升,且在较高误码率即10-1时依然可以保持稳定的识别效果。In the face of the transmission characteristics of high bit error rate in space-based networks,and to ensure the efficient and reliable transmission of data between heterogeneous networks,a recognition system method based on integrated learning for AOS(Advanced Orbiting Systems)protocol in spatial link layer is designed.This method uses the integrated learning model to learn the AOS protocol data,constructs the AOS protocol recognition model based on the integrated learning,realizes the AOS protocol recognition accurately,and carries on the experimental verification under the high bit error rate.The experimental results show that the integrated learning model has a good recognition effect in AOS protocol recognition,the recognition efficiency has been improved significantly,and it can still maintain a stable recognition effect at a high bit error rate of 10-1.

关 键 词：机器学习 集成学习 空间链路层协议 +高级在轨系统协议 协议识别 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]