Study Finds Autonomous Vehicles Safer in Most Scenarios but with Challenges in Low Light and Turning

TapTechNews June 19th news, local time on the 18th, Nature Communications published the latest technical research titled Matched case-control analysis of autonomous vehicle and human-driven vehicle accidents.

The study said that autonomous vehicles have fewer accidents in most scenarios than human-driven vehicles, except under insufficient light and when turning - at this time, autonomous driving is more difficult to handle than human driving.

The above conclusion was put forward by two researchers from the University of Central Florida, Mohamed Abdel-Aty and Ding Shengxuan (transliteration, Shengxuan Ding). The two collected 2,100 accident data from California and the National Highway Traffic Safety Administration (NHTSA) which involved vehicles equipped with some degree of autonomous driving or driving assistance technology. They also collected more than 35,000 accident data involving unassisted manual driving.

Two researchers used a statistical matching method to find accidents that occurred under similar circumstances (TapTechNews note: Similar factors include road conditions, weather, time and the location of the accident, such as at an intersection or a straight road).

Abdel-Aty said that the overall results show that autonomous cars generally show higher safety in most cases. But the analysis also found that autonomous cars have a five times higher risk of collision at dawn and dusk than manual driving, and the chance of having an accident almost twice as high as manual driving when turning.

Ding Shengxuan said that it is very important to improve the safety of autonomous cars under dawn and dusk or turning conditions, The key strategies include enhancing weather and lighting sensors, and effectively integrating sensor data.

TapTechNews attached research report: https://www.nature.com/articles/s41467-024-48526-4