On the Use of Differential Correction Clustering for Facing Spoofing Attacks to GNSS Augmentation Networks

Nowadays, Global Navigation Satellite Systems are the main source for high accuracy positioning and timing. For this reason, they are essential both for everyday activities and services, and for the industrial and critical infrastructure sectors. Moreover, the spread of increasingly autonomous vehicles results in strict accuracy and integrity requirements. This leads to the need for additional infrastructure to send corrections to the end users and mitigate the measurement errors, the Augmentation Networks. However, due to the increasing exploitation of localization functionalities, the Augmentation Networks could become a primary target for attackers resulting in a high financial and safety cost. Among the possible attacks, spoofing, that is the generation of a fake satellite signal which is seen as genuine by the receiver, is one of the most powerful and tricky. In this contribution, a detection and mitigation strategy for Augmentation Network spoofing attacks is proposed. We introduce two attack models and present a technique based on K-means clustering to counteract them. More in details, our approach is based on the computation of the number of clusters formed by the Augmentation Network corrections. Starting from the hypothesis that under nominal conditions only one cluster is present, the effects of the attacks on the clustering procedure are analyzed, and several attack simulations are performed to evaluate the algorithm performances. The proposed method has been compared both to an Augmentation Network attack detection technique, and to a receiver-level spoofing mitigation approach, showing comparable or better performances. Moreover, to the best of our knowledge, this is the first work addressing mitigation for spoofing attacks which target an Augmentation Network.