An Exploratory Analysis of Crashes Involving Autonomous Vehicles

Autonomous vehicles (AVs) have the potential to lower the number of vehicular crashes, as research has shown that most collisions are caused by driver error. Consequently, several companies in the US and other countries have conducted pilot tests on public highways to expedite the mainstream deployment of AVs and examine the effects that they have on traffic safety. Although a few studies have used AV crash data to attempt this task, their sample sizes were too small, and they failed to explore many of the factors causing the crashes. This study aims to analyze the characteristics of the collisions in terms of frequency per annum; company and make of AV; and vehicular, environmental, and roadway factors. The crash data, gathered for this research from the California Department of Motor Vehicles, pertains to the characteristics of 259 collisions involving AVs that occurred in California between September 2014 and June 2020 and reveals a sharp increase in the number of crashes after 2017. AVs were at fault for a small proportion of them, as most of them were rear-end collisions that occurred when they were operating in the autonomous mode, and the number of passengers sustaining a major injury was very low. Environmental factors such as weather and lightening did not have a significant impact, as the data shows that most of the accidents occurred in clear weather during daylight hours. The results of this study will help transportation experts more fully understand the advantages and disadvantages of autonomous vehicles by providing a comprehensive picture of collisions involving them.