Revolutionizing Road Safety: Autonomous Vehicles and Human Error in Traffic Incidents

A study by US researchers published in the journal Nature Communications has found that autonomous vehicles are generally less likely to cause accidents compared to those driven by humans. However, there are exceptions to this trend, particularly when it comes to accidents at dusk and while turning.

Mohamed Abdel-Aty and Shengxuan Ding from the University of Central Florida conducted the research, evaluating over 37,000 accidents, including 2,100 involving highly or partially automated vehicles. While partially automated cars require someone to sit behind the wheel and control assistance systems, fully autonomous vehicles do not require any human intervention.

The study revealed that there has been progress in terms of safety with autonomous vehicles, but limited accident data makes it difficult to determine their exact differences in risk compared to human-driven vehicles. The researchers analyzed information on accidents involving autonomous vehicles from various American databases and compared them with over 35,000 accidents involving vehicles driven by humans. They found that pedestrian accidents were significantly lower for autonomous vehicles compared to human drivers.

Furthermore, the researchers conducted a special analysis to consider factors such as traffic volume, weather, road surface, and location to make predictions about accident probability in different situations. The study showed that highly automated vehicles have a lower accident probability in rain compared to those driven by humans. However, they also found a higher risk of accidents with autonomous vehicles in challenging visibility conditions at dusk and while turning. Overall, the researchers conclude that highly automated driving systems are safer than human-driven vehicles in most scenarios due to their object detection and avoidance capabilities.

In conclusion, while there have been significant advancements in technology with regard to self-driving cars, it is crucial that we continue to monitor their performance under different scenarios before we can fully rely on them for our daily transportation needs.