河北大学学报(自然科学版) ›› 2019, Vol. 39 ›› Issue (3): 323-329.DOI: 10.3969/j.issn.1000-1565.2019.03.015

• • 上一篇    

基于SVDD的ADS-B异常数据检测

王振昊,王布宏   

  • 收稿日期:2018-10-16 出版日期:2019-05-25 发布日期:2019-05-25
  • 作者简介:王振昊(1993—),男,山东冠县人,空军工程大学在读硕士研究生,主要从事传感器数据融合研究. E-mail:472070767@qq.com
  • 基金资助:
    国家自然科学基金资助项目(61671465)

ADS-B anomaly data detection based on SVDD

WANG Zhenhao, WANG Buhong   

  1. College of Information and Navigation, Air Force Engineering University, Xian 710077, China
  • Received:2018-10-16 Online:2019-05-25 Published:2019-05-25

摘要: 自动相关监视广播(ADS-B)与空管二次监视雷达(SSR)是空中交通管制的2种重要监视手段,其中ADS-B是目前正在推出的通信协议,在下一代空管监视系统中将会发挥重要作用.然而,ADS-B协议中安全措施缺乏,容易遭受虚假数据注入的攻击.为了识别ADS-B中的异常数据,利用与其同步的SSR数据及通过Kalman滤波得出的的协方差矩阵,得到一组具有多维属性特征的样本数据,使用支持向量数据域描述的方法(SVDD)训练样本数据,可以得到用于检测异常的分类器.利用此分类器检测之后收到的ADS-B数据,从而识别出异常数据.通过仿真实验表明,该方法对于ADS-B异常数据具有80%以上的正确识别率,其中对于固定偏差注入的检测虚警率为5%,漏警率为0,对于随机偏差注入的检测虚警率为5%,漏警率为12.5%,验证了该方法的可行性.

关键词: 自动相关监视广播, 空管二次监视雷达, 支持向量数据域描述, 虚假数据注入, 异常检测

Abstract: The Automatic Dependent Surveillance Broadcast(ADS-B)and the Secondary Surveillance Radar(SSR)are two important monitoring methods for air traffic control. ADS-B is the communication protocol currently being introduced, which will play an important role in the next generation of air traffic surveillance. However, the ADS-B protocol lacks security measures and is vulnerable to false data injection attacks. In order to identify the anomaly data in ADS-B, a set of sample data with multi-dimensional attribute features is obtained by using the SSR data synchronized with it and the covariance matrix obtained by Kalman filtering, using the method of support vector data field description(SVDD). By training the sample data, a classifier for detecting anomalies can be obtained. This classifier is used to detect the ADS-B data received later, thereby identifying abnormal data. The simulation experiment shows that the method has a correct recognition rate of more than 80% for ADS-B anomaly data, and the detection false alarm rate for fixed deviation injection is 5%, the false alarm rate is 0, and the false alarm is detected for random deviation injection. The rate is 5% and the missed alarm rate is 12.5%, which verifies the feasibility of the method.

Key words: automatic dependent surveillance broadcast, secondary surveillance radar, support vector data description, false data injection, anomaly detection

中图分类号: