Machine Learning Techniques for Anomaly Detection in Autonomous Vehicles: a review

Smart vehicles have been designed to reduce human-error accidents. However, these vehicles, equipped with advanced technologies and wireless communication, are vulnerable to cyberattacks. While ۹۰% of current accidents are caused by human error, cyberattacks pose a new threat to road safety. Therefore, the security of smart vehicles has become a significant challenge. The objective of this paper is to present a comprehensive review of the solutions proposed by researchers in the field of intelligent vehicles for the detection of anomalies in in-vehicle networks based on machine learning techniques. Initially, we will examine the work conducted in the area of anomaly detection in autonomous vehicles from ۲۰۲۰ to ۲۰۲۴, describing the type of technique employed in each study. Subsequently, in the final section, we will provide a comparison of various dimensions of the conducted research, based on the type of dataset used, the algorithm employed, and the detected attacks.

کلیدواژه ها:

Connected Autonomous Vehicle(CAV) ، Machine Learning ، Intrusion Detection System(IDS) ، CAN