Human drivers have enough trouble in bad weather: we need to look no further than the most recent massive winter storm–Jonas, this past January–for proof. Jonas made the entire East Coast grind to a halt, with every news channel showing images of cars and trucks lined up in ditches, half-buried in snow. Some cities even banned car travel. In Washington, D.C., drivers abandoned their vehicles, choosing instead to walk (and racking up a city-wide parking bill of over $1 million), rather than force their cars to attempt a venture through the elements.
Since mass-produced cars first rolled off assembly lines over 100 years ago, the auto industry has been working to improve vehicle control in inclement weather conditions, and the latest advancements in vehicle technology—particularly driverless vehicles—are no exception.
One of the main selling points of driverless vehicles is safety: as we’ve previously noted, human error accounts for 94 percent of all traffic crashes. But bad weather throws a wrench in driverless vehicles’ safety claims because, as several outlets report, the unpredictable natural elements are tough for driverless vehicles to navigate, too. This presents another hurdle the autonomous vehicle industry will have to overcome before the cars are introduced to consumers.
We’ve talked about many driverless vehicle concerns at Quoted, from philosophical and psychological concerns to the always-critical safety considerations, but with about 70 percent of U.S. residents living in the so-called snow belt, the ability of driverless vehicles to navigate poor weather conditions has truly become a key issue.
How “Snow Smoke” and Other Weather Impairs Driverless Vehicles
Bad weather such as snow, ice, rain, and even fog can interfere with how driverless vehicles see and interpret their world. Driverless technology uses cameras, radar, and lidar sensors (which bounce laser light off objects) to, as Bloomberg Business reports, “provide 360-degree detection of lanes, traffic, pedestrians, signs, stoplights and anything else in the vehicle’s path. That enables it to decide, in real time, where to go.” Snow can block cameras and limit LIDAR range because snow can misdirect light reflection, “potentially confusing a curtain of falling snow with something to avoid, causing the vehicle to hit the brakes.”
“Snow smoke” is also a term that might become more common as driverless technology becomes mainstream. Frozen flakes of snow, kicked up from the wheels, can accumulate on any type of vehicle, but with a driverless one, the icy buildup can block sensors, disabling the autonomous technology and rendering the car unable to navigate the road.
These issues must of course be overcome if autonomous vehicles ever hope to become mainstream.
Can Driverless Cars Adapt for All Weather Conditions?
In January, Ford issued a press release claiming they are testing a fully autonomous vehicle in winter weather conditions. Ford’s driverless technology, writes Bloomberg Business, “scans roads in advance with lidar to create high-definition 3-D maps that are much more accurate than images from global-positioning satellites, which can be 10 meters (33 feet) off.”
Jim McBride, Ford’s technical leader for autonomous vehicles, told The Daily Mail that the goal for autonomous vehicles is to eventually be able to detect and assess driving conditions, adjusting and even stopping when necessary.
Google is already practicing this advanced detection. The Daily Mail reports that the tech giant uses laser sensors with windshield wipers to navigate poor visibility conditions like rain and exhaust clouds. For now, Google’s self-driving cars pull to the side of the road and wait out storms, but they are learning to navigate bad conditions and adjust just as a human driver would.
Sometimes simply moving sensors to different locations on the vehicle can make a big difference in the car’s ability to drive in poor conditions. Volvo engineers moved their radars behind the windshield in their latest autonomous model—the XC90—Bloomberg Business reports, and now wipers can easily clear away snow and ice.
Solution: Make Driverless Cars More “Human”?
Even Jaguar Land Rover, who just this fall publicly stated they were declining to enter the driverless vehicle race, is working on autonomous technology. The British automaker has joined a new research project called MOVE-UK, Live Science reports, which will work to develop safer, more effective driverless vehicles by studying how human drivers navigate situations like poor driving conditions. The goal of the project is to develop driverless cars that perform like people, not robots, which, they hope, will make consumers more comfortable and will improve safety standards.
The ultimate solution, says Bloomberg Business, might be found in artificial intelligence (AI). Danny Shapiro, senior director of automotive at Nvidia Corp, a California-based supplier of high-speed processors, explains that their latest “computer brain” can perform “24 trillion ‘deep learning operations’ per second.” Deep learning could translate into an autonomous vehicle that’s able to navigate road conditions based on millions of miles of driving experience researchers upload to the vehicle’s software, and then constantly update.
While AI is still being developed, major carmakers are now testing their driverless technology in inclement weather. Volvo in Sweden (in a town just north of the Arctic Circle), Google in Lake Tahoe, and Ford in Michigan are just a few of a growing list of automakers working to make autonomous vehicles that are prepared for any weather conditions.
Ryan Eustice, an associate professor of engineering at the University of Michigan who is working with Ford on snow testing, told Bloomberg Business: “A car that’s able to do nationwide, all-weather driving, under all conditions, that’s still the Holy Grail.”