Eine neue Studie prognostiziert, dass vollautomatische Fahrzeuge oder „selbstfahrende Autos“ in den nächsten zehn Jahren die Zahl der Verkehrsunfälle in den USA verringern werden. Das konservativste Szenario einer Einführung von 1 % prognostizierte 67.408 vermiedene Verletzungen. Das optimistischste Szenario einer Einführung um 10 % prognostizierte einen Rückgang von 1.078.528 Verletzungen.

Fully automated vehicles and road safety: Potential for reduction of traffic collisions

A Sunnybrook-led **study predicts that fully automated vehicles, or ‘self-driving cars’, will reduce road traffic collisions in the United States over the next ten years** across various adoption and road safety scenarios.

The study was published today in JAMA Surgery.

Investigators predicted collisions from 2025 to 2035 under various automated vehicle adoption scenarios – ranging from one per cent to 10 per cent – using data from the National Highway Traffic Safety Administration and Federal Highway Administration in the United States.

**The study’s most conservative scenario of one per cent adoption predicted 67,408 avoided injuries, translating to a 0.2 per cent reduction in expected injuries. ​The most optimistic scenario of ten per cent adoption forecasted a reduction of 1,078,528 injuries, representing a 3.6 per cent decrease**.

“There’s strong potential for self-driving cars to significantly reduce traffic collisions,” says Dr. Armaan Malhotra, neurosurgery resident physician and the study’s co-investigator. “Currently, highway driving is associated with a much higher rate of collisions, versus driving in a city or urban setting. And we know that in the United States, self-driving cars are primarily used in city-settings. It will be important to study the trends as more fully automated cars become available in different scenarios. Evolving technology will also have an impact on safety. Ultimately, data will need to guide policymakers and regulators as fully automated vehicles become commonplace on public roads.”

Driver behavior like distraction, impairment and speeding are major causes of collisions. Higher levels of automation in cars have the potential to mitigate the contribution of risky and dangerous driver behaviors to national road traffic injuries. This work represents a critical step forward in understanding the potential public health impacts of scaling fully automated vehicles across North America.

For those interested, here’s the link to the peer reviewed journal article:

https://jamanetwork.com/journals/jamasurgery/article-abstract/2843251

..assuming self-driving cars never go wrong.

So I get that human error is egregious, but in the case of 1% up to 10% adoption wouldn’t collisions either stay the same or even go up because the 90-99% of human drivers will be driving poorly as they normal do and crash into the self driving cars because they aren’t used to reacting to the driving style of self driving cars? As well as get impatient with them driving slower than human drivers and be overtaking them more often thus creating more opportunities for them to crash?