Connected and Autonomous Vehicles (CAVs) will play a critical role in improving the safety and efficiency of future Intelligent Transportation System (ITS). Periodic broadcasts of basic safety messages (BSMs) containing up-to-date vehicle status information constitute one important class of inter-vehicular communication. If false or inaccurate information is inserted in the BSMs it can lead to serious consequences such as accidents resulting in bodily injury and even loss of life. In this paper, we propose a machine learning based approach for automatically detecting replay attacks, where BSMs received from neighboring vehicles are rebroadcast, with false sender information, by a malicious node. Simulations using the publicly available Vehicular Reference Misbehavior (VeReMi) extension dataset demonstrate that the proposed model clearly outperforms existing techniques for detecting BSM replay attacks.

A Proof-of-Location (POL) system is used to issue a proof-of-location token ($pol$) to a user who has been present at a location$\ell oc$, such that it can be later presented to a verifier to assure the presence of the user at$\ell oc$. Basic POL security requirements areunforgeabilityof$pol$, and itsnon-transferability(a$pol$issued to user$u_1$cannot be used by$u_2$). An additional important property of POL systems isuser privacyagainst the issuers and verifiers. We make two contributions. First, we formalize the POL security and privacy properties, and construct the first system providing provable security and privacy against the issuer and the verifier, both. Second, we introduce ageo-tampering attackthat completely breaks POL system security, by simply changing the location of a$pol$issuing node. The attack applies to portable infrastructure nodes that are not continually monitored. We propose an algorithm that is used by a$pol$issuer to provide a location integrity “proof”, that will be embedded in a$pol$to protect against this attack. The proof relies on a novel application of euclidean Distance Matrices. We implemented our POL on an off-the-shelf Android smartphone to show the practicality of the proposed algorithms.

Although connected mymargin autonomous vehicles (CAVs) hold great potential to improve driving safety and experience significantly, cybersecurity remains a critical concern. As the de-facto standard for in-vehicle networks, the Controller Area Network (CAN) carries messages and commands vital to the operation of the vehicle. However, due to a lack of security mechanisms, intruders are able to conduct devastating attacks on drivers and passengers over CAN. In order to safeguard CAVs, an Intrusion Detection System (IDS) can be deployed to monitor CAN network activities and detect suspicious behavior resulting from an attack. This paper proposes a prediction-based IDS framework for detecting anomalies and attacks on a CAN bus using temporal correlation of message contents. Two candidates are introduced as the prediction module. The first network is an LSTM that predicts time series data separately for each CAN ID, and the second is a ConvLSTM that predicts messages using correlated data of several CAN IDs. An attack is classified according to prediction errors by a Gaussian Naïve Bayes classifier. The proposed IDS is evaluated against other state-of-the-art one-class classifiers, including OCSVM, Isolation Forest, and Autoencoder, and three existing works, including ReducedInception-ResNet, NeuroCAN, and CANLite, using a real-world dataset, the Car Hacking Dataset. A comparison between the two suggested architectures and their use cases is given. Compared to baseline methods and related studies, the proposed method is shown to be more accurate and can achieve F-scores and detection accuracy of almost 100%.