Enabling AI toward future 6G vehicular network security
The fifth-generation (5G) network is rapidly deployed around the world, after which artificial intelligence-enabled next-generation (6G) will be in line with the future evolution of network intelligentization . Proceedings of the IEEE’s invited paper “Future Intelligent and Secure Vehicular Network Toward 6G: Machine-Learning approaches” published in February 2020 discussed the role of machine learning for vehicular network intelligentization in 6G. The article discussed challenges and applications of vehicular communication, networking, and network security and their relation to machine learning for 6G applications.
In general, for vehicular network security, there are challenges related to high dynamics of vehicle nodes, heterogeneous structures with various devices and technologies, increasing scale of the vehicular network, and privacy requirements. The two primary security architectures to ensure vehicular communication’s confidentiality adopt the vehicular public key infrastructure (PKI) and elliptic curve digital signature algorithm (ECDSA). However, malicious attacks, including Sybil attack, denial of service (DoS), blackhole attack, wormhole attack, the bogus information attack, and replay attack capable of forging identity, hijacking normal nodes can also break through the PKI and ECDSA security. To detect and prevent such intrusions, many malicious behavior detection methods have been proposed in the past. Machine learning-based data mining approaches have shown promising performance for intrusion detection in vehicular networks in recent days. Both shallow learning and deep learning techniques have been studied for misuse and anomaly detection for attacks. Since misuse-based detection algorithms depend on the existing signatures of attacks, they cannot detect novel attacks. In contrast, anomaly-based detection algorithms model behavior as a benchmark and identifies anomalies as deviations from the normal behavior for which they are robust in detecting some novel attacks while handling the changing environment in real-time. In a practical setting, both existing and novel attacks can occur. Such phenomena require a combination of both misuse and anomaly detection as a hybrid mechanism where some detected results are labeled with signatures for misuse detection besides the signature-less anomaly detection. A hybrid mechanism can save both human resources and time for detection. Moreover, there are some proactive security approaches leveraging reinforcement learning and imitative learning to improve proactive detection performance.
The 6G network aims to develop a highly dynamic and intelligent system with an adaptive network to support various application requirements and service types. Thus, the 6G network also requires more strict constraints of security for global intelligentization. From the communication perspective, confidentiality, integrity, authenticity, and availability need to be satisfied. In the traditional vehicular network, enhancing security features are often linked to increasing communication costs. In particular, proactive security solutions with exploration-based detection have high communication costs. Thus, the existing security approaches make a tradeoff by relying on more reactive detection to save the communication cost while providing the necessary security solution. The communication resources in the 6G network will be unlimited with ultra-reliable and low-latency communication (uRLLC) and massive machine-type communications (mMTC), and enhanced mobile broadband (eMBB). Therefore, it can be anticipated that 6G will be the new paradigm for proactive and exploration-based security approaches.
 K. B. Letaief, W. Chen, Y. Shi, J. Zhang, & Y. J. A. Zhang, The roadmap to 6G: AI empowered wireless networks. IEEE Communications Magazine, 57(8), 84-90, 2019.
Written by Nishat Mowla
Earlier this year in an AutoSec Newsletter it was reported that source code of mobile apps and internal tools developed and used by Nissan North America was circulating the Internet. The data had been shared on Telegram channels and different forums for hackers in the...
Tracker Network (UK) Ltd is a company that provides vehicle systems used for recovering stolen vehicles in the UK, and its stolen vehicle systems operate in all 52 police forces in UK. Tracker’s tracking solution combines GSM, VHF and GPS which available localisation...
In modern vehicles, powertrain design and control, traffic flow, energy demand, and emissions strongly influence each other. Along the way, control algorithms will leverage information from vehicle environment enabled by new communication technologies. In March 2021,...