Driverless cars are getting better at navigating roads each day, but there's a caveat: they have to be well-paved, clearly marked roads. And any regular driver can attest that is just not a reasonable thing to expect in America.
Reuters shed light on the issue when Lex Kerssemakers, CEO of Volvo Cars of North America, was angry his semi-autonomous car couldn't drive itself during a press event at the Los Angeles Auto Show because of the road's poor quality.
"It can't find the lane markings!" Reuters reported Kerssemakers saying to Mayor Eric Garcetti, who was at the wheel. "You need to paint the bloody roads here!"
An estimated 65% of US roads are in poor condition, according to a 2014 White House report on transportation infrastructure. The report adds that in the last decade the US has fallen from 7th to 18th overall when it comes to road quality, according to the most recent World Economic Forum.
To address the issue, automakers have begun mapping routes so cars know the road layouts and traffic rules in advance.
Audi, Daimler, and BMW acquired Here, the former mapping division of Nokia, to make real-time maps to aid their driverless car efforts and Google is also mapping its own routes. Toyota also announced in December it would begin drawing information from customers' camera-equipped vehicles to gather road images and vehicle positional information.
Google-turned-Alphabet has also teamed up with the Department of Transportation to aggregate anonymized data to better understand traffic patterns and congestion areas in cities that will make it easier for driverless cars to navigate cities.
The project, called Flow, will use data from billions of miles of trips, Waze, Google Maps, and sensors on the roads to understand transportation patterns.
Andrew Ng, chief scientists at Baidu, often referred to as the Chinese version of Google, has written about this very issue in Wired. To get driverless cars on the road, it is necessary to make "modest changes to our infrastructure," he wrote in the Wired article.
"Safe autonomous cars will require modest infrastructure changes, designs that make them easily recognized and predictable, and that pedestrians and human drivers understand how computer driven cars behave," Ng wrote.
