Inclement Weather Poses Big Obstacle For Autonomous Vehicle Researchers
Print this Article | Send to Colleague
The first semi-autonomous vehicles are set to reach showrooms within the next several years, and several automakers are promising they’ll have the first fully-autonomous products on the road by 2020. That timetable might work in some parts of the country. But if you live in snowier climes, don’t be surprised if you’re in for a much longer wait, especially if you’re hoping to have your car take over driving duties in inclement weather.
Google recently began testing what will soon be a fleet of at least 100 self-driving prototypes near its Silicon Valley headquarters. Most of those pilot vehicles won’t even be equipped with steering wheels and pedals, just a microphone where passengers can input a destination.
"The Google car is a good show car, but it’s not going to be something you can drive on (Detroit’s) I-75 when it’s snowing," cautions Steffen Linkenbach, Director of the Systems and Technologies group for automotive supplier Continental North America. "It will be unacceptable to customers that the system won’t work because it’s snowing or the sun is shining at the wrong angle," he adds.
While driver error is blamed for ninety percent of all collisions,
humans still can do a much better job at estimating where to point their
vehicles on a snow-covered road, where sensor technologies currently
have to give up because they can’t make out lane markers. They also can
be fooled by streaks of salt left on dry pavement from the last storm,
and even by bright sun glaring off the road at the wrong angle.
One of the largest suppliers of key driver assistance components, Continental has become one of the leaders in autonomous vehicle research. But it is pushing back on the aggressive timetables some of its automaker clients have laid out for putting self-driving technology into production.
The most basic driver assistance systems rely on wheel and yaw sensors to get a sense of what’s happening to a vehicle. More advanced cars now add a blend of radar, sonar, cameras and an advanced laser technology known as LIDAR.
Bringing them all together is an increasingly complex task. Even today’s more advanced vehicles, like the S-Class, use dozens of times more software code than an F-16 fighter jet. Fully autonomous systems will bump that up by several orders of magnitude.
Back to NAFA Connection
