基于K-means聚类的GPS同步式欺骗识别方法

针对进入GPS接收机的高隐蔽性同步式欺骗信号和真实卫星信号的区分问题，该文提出一种基于多重基带特征进行聚类分析的欺骗干扰识别方法。首先总结了抗同步式欺骗的基带处理流程；然后在欺骗检测和估计的基础上，在信号处理层级提取各个信号分量的伪码和载波多普勒频率一致性(CCDC)特征、相对幅度关系特征、载噪比异动特征，在信息解算层级提取各信号分量和其他信号分量相组合进行解算对应的伪距残差特征，以及欺骗入侵前后的钟差异动特征；最后利用K均值聚类(K-means)算法对不同特征进行综合，从而完成对欺骗信号的识别。基于射频(RF)信号采集回放的半实物实验及反欺骗软件接收机的处理结果表明，该方法能够及时而准确地识别同步式欺骗和真实信号，从而指导接收机抑制欺骗干扰并恢复正确的解算结果。

Synchronous GPS Spoofing Identification Based on K-means Clustering

To solve the problem of distinguishing highly-concealed synchronous spoofing signals from authentic signals inside a GPS receiver, a spoofing identification method based on clustering analysis of multiple characteristics is proposed. The procedure of baseband anti-spoofing processing is summarized. On the basis of spoofing detection and signal parameter estimation, the Code-Carrier Doppler frequency Coherence (CCDC), relative amplitude, and carrier-to-noise ratio variation characteristics are extracted as the spoofing identification characteristics at the signal processing level. At the information processing level, the pseudo-range residual corresponding to the combination of different signal components and the receiver clock variation before and after the spoofing attack are calculated. Then the K-means clustering is used to identify the spoofing signal combining these characteristics. The semi-physical experiment based on Radio Frequency (RF) signal sampling and playback and the processing results of our anti-spoofing software receiver demonstrate that the proposed method can identify synchronous spoofing timely and accurately, thus guiding the receiver to mitigate the effects of spoofing and recover correct navigation solutions.