基于深度学习的ADS-B异常数据检测模型

广播式自动相关监视（ADS-B）是下一代空中交通运输系统的重要组成部分，是新航行系统中非常重要的通信和监视技术，但其协议没有提供相关的信息认证和数据加密，因此极其容易受到欺骗干扰的影响。针对ADS-B报文数据特点，采用深度学习的seq2seq模型对ADS-B报文数据进行重构，通过重构误差来检测异常，并对数据进行特征扩展，使模型能更好的捕捉数据的时间依赖性。实验结果表明，所采用的方法优于传统的机器学习方法，且在数据特征扩展后，模型检测效果提升。相比于现有的欺骗干扰检测方法，该方法不需要改变ADS-B系统的协议，也不需要额外的节点或传感器参与，具有一定的适应性和灵活性。

ADS-B anomaly data detection model based on deep learning

Automatic Dependent Surveillance-Broadcast (ADS-B) is an important part of the next generation air transportation system. It is a critical communication and monitoring technology in the new navigation system, but its protocol does not provide relevant authentication and data encryption, so it is extremely vulnerable to various spoofing attack. Based on the data characteristics, this paper uses the deep learning seq2seq model to reconstruct the ADS-B time series, and the reconstruction error can detect the anomalous ADS-B messages. Extending the feature space of time series enables the model to better capture the time dependence to further improve the effect of anomaly detection. The experimental results show that the proposed method is superior to traditional machine learning methods and time series enrichment can improve detection results. Compared with the existing spoofing attack detection method, the proposed method does not need to change the ADS-B protocol and does not require additional participating nodes or sensors, and has certain adaptability and flexibility.