Machine Learning Technologies for Secure Vehicular Communication in Internet of Vehicles: Recent Advances and Applications

Table 1

Summary of artificial intelligence methods in secure Vehicle-to-Everything networks.

Year	Source	Security approaches	Features	Advantages	Challenges	Citations
2020	ArXiv	NSL-KDD data mining; Cloud Security Alliance (CSA)	Machine learning in fifth generation (5G) IoV	Security issues related to softwarization, software-defined perimeter, and virtualization	QoS performance and scalability and cost in secure V2X dynamic networks	Abdallah [20]
2020	Elsevier	Controller Area Network (CAN); IDS; Security-Aware FlexRay Scheduling Engine (SAFE); Hardware Security Module (HSM)	AI-based V2X automotive security framework	Detects sensing and communication layers’ attacks	Cybersecurity in fully autonomous V2X	El-Rewini [21]
2019	arXiv	Intelligent V2X security (IV2XS); physical layer security (PLS)	Cognitive security based on context-aware proactive security	Security decision-making according to vehicles’ channel conditions	Identify the best-suited level of security.	Furqan [22]
2019	WiSec’19	Basic safety messages (BSMs).	Misbehavior detection based on ML for secure V2X traffic	Detects spoofing attacks in the V2X application layer	Identify and detect the V2X location spoofing	So [24]
2018	IEEE	MinMax, MLP, Adaboost, and Random Forest misbehaving classifiers	V2X traffic safety-based ML algorithms	A misbehavior classifier for vehicle data classification	Secure decision for V2X traffic safety	Monteuuis [26]
2016	PLoS ONE	Controller Area Network (CAN) and IDS	Intrusion detection system (IDS) based on deep neural network (DNN)	Extract the statistical properties of normal and attack CAN data packets	Identify malicious attack to V2X networks	Kang [27]