Enhancing Cybersecurity and Privacy Protection for Cloud Computing-Assisted Vehicular Network of Autonomous Electric Vehicles: Applications of Machine Learning

Tiansheng Yang, Ruikai Sun, Rajkumar Singh Rathore, Imran Baig

Due to developments in vehicle engineering and communication technologies, vehicular networks have become an attractive and feasible solution for the future of electric, autonomous, and connected vehicles. Electric autonomous vehicles will require more data, computing resources, and communication capabilities to support them. The combination of vehicles, the Internet, and cloud computing together to form vehicular cloud computing (VCC), vehicular edge computing (VEC), and vehicular fog computing (VFC) can facilitate the development of electric autonomous vehicles. However, more connected and engaged nodes also increase the system’s vulnerability to cybersecurity and privacy breaches. Various security and privacy challenges in vehicular cloud computing and its variants (VEC, VFC) can be efficiently tackled using machine learning (ML). In this paper, we adopt a semi-systematic literature review to select 85 articles related to the application of ML for cybersecurity and privacy protection based on VCC. They were categorized into four research themes: intrusion detection system, anomaly vehicle detection, task offloading security and privacy, and privacy protection. A list of suitable ML algorithms and their strengths and weaknesses is summarized according to the characteristics of each research topic. The performance of different ML algorithms in the literature is also collated and compared. Finally, the paper discusses the challenges and future research directions of ML algorithms when applied to vehicular cloud computing.