The research supports the view that autonomous technology could improve road safety, while also identifying specific areas for improvement.
Autonomous vehicles, with their sophisticated sensors and algorithms, are designed to navigate traffic with precision, potentially leading to safer driving environments by reducing human error, which is a leading cause of accidents on the road today.
However, understanding the conditions under which autonomous vehicles perform better or worse than human drivers is crucial for maximising their safety benefits.
Mohamed Abdel-Aty and Shengxuan Ding conducted an analysis comparing accident data (mainly in California, United States) collected from 2100 autonomous vehicles and 35,133 human-driven vehicles from between 2016 and 2022. The authors found that autonomous vehicles were generally safer and less likely to be involved in accidents when performing routine driving tasks, such as maintaining lane positions and adjusting to the flow of traffic. Additionally, autonomous vehicles were also shown to be safer upon rear-end and side-swipe accidents, which were 0.5 and 0.2 times less common, respectively.
However, the data shows that autonomous vehicles seem to be more prone to accidents in specific situations, such as during low-light conditions at dawn or dusk and while executing turns (5.25 and 1.98 times higher than human-driven vehicles, respectively).
The authors noted that those were areas where autonomous driving technology needed further refinement to match or exceed human driving capabilities consistently.
Although the study suggested that autonomous vehicles had the potential to improve road safety, it also underscored the importance of addressing technological limitations to ensure they can function safely in a broad range of circumstances.
The authors note that there were limitations to their findings.
In particular, with regards to current and emerging autonomous vehicle technology levels as well as the specific systems that activate during accidents.
They concluded further research was needed to provide a more nuanced understanding of safety differences between autonomous and human-driven vehicles.
