Anomaly Detection for Connected and Automated Vehicles: Accident Analysis

Abstract

Smart mobility is a key component of smart cities, and the switch from traditional automotive systems to connected and automated vehicles (CAVs) is recognized as one of the evolving technologies on urban roads. Although the current autonomous vehicle (AV) mobility environment may be geared toward infrastructure and road users, it cannot facilitate the adoption of CAV in the future due to the presence of different modules that are nested in the cyberspace. Furthermore, the ability to make accurate decisions in real-time is essential for the success of autonomous systems. However, cyberattacks on these entities might skew the decision-making processes, which can result in complex CAV mishaps. Furthermore, the method utilized by the police to conduct accident investigations cannot be used to identify road accidents brought on by cyberattacks. Therefore, this paper proposes a 5Ws & 1H-based investigation approach to deal with cyberattack-related accidents. Also, a stochastic anomaly detection system is proposed to identify the abnormal activities of the automated driving system (ADS) functions during a crash analysis. Further, two case studies are shown to validate the results of the proposed algorithms.