Communications of the ACM
Will Driverless Cars Really Save Millions of Lives? Lack of Data Makes It Hard to Know
A lack of data on how driverless automobiles' performance compares with human drivers makes it difficult to quantify the safety benefits of autonomous vehicles.
Credit: Tony Avelar/AP
The lack of data on how driverless automobiles' performance compares with human drivers makes it difficult to determine the safety benefits of autonomous vehicles.
Google's Chris Urmson says part of the problem "is we don't actually have a good understanding of how good human drivers really are." The U.S. government does not maintain a comprehensive database on automobile crashes, which makes determining whether computer- or human-driven cars are more likely to crash problematic.
Google wants the U.S. National Highway Traffic Safety Administration (NHTSA) to produce a comprehensive database, while NHTSA chief Mark Rosekind warns automated vehicles are no safer than cars driven by humans. Rosekind has asked manufacturers, software developers, and safety experts across the U.S. to devise new safety metrics for the realities of automated driving.
Meanwhile, skeptics of automated car technology are concerned with federal officials recommending the government obtain more insights on effective safety metrics and testing standards from industry. Industry and federal officials often cite the statistic of human error causing 94% of U.S. car crashes as the key motivator for putting driverless cars on the road.
However, tracking actual safety gains will be complicated because for years humans will be superior to robots in different areas, and vice versa.
From The Washington Post
View Full Article - May Require Free Registration
Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA
No entries found