A modern car is equipped with many Electronic Control Units (ECUs) with connections to external networks. As software takes significant portions of vehicles, vulnerabilities in the software and in-vehicle network protocols may increase risk of car accidents. Although many defense mechanisms such as Intrusion Detection Systems (IDSs) have been proposed for the vehicle, these IDSs have limitations in investigating causes of the car accident. Digital forensics for the vehicles have not been thoroughly studied yet. In this paper, we propose algorithms providing comprehensive information derived from limited vehicle data. Also, our proposed algorithms can be used for investigating a car accident, including cyber-attack. We evaluate the performance of the proposed algorithms using simulated vehicle data.
Lee et al. (2026) studied this question.