Automakers are working on driverless car projects that would improve robotic vehicles' capacity to handle chaos driving conditions like snow as well as a human.
One of the situations most difficult to handle for a self-driving car is snow. For people living in the north east of the United States, an autonomous car that would not be able to operate safely in snow would be useless for many months per year.
According to the publication Cars, even the clearest and best-marked roads with bright lines can become a problem when a layer of snow covers them. A human can still drive, but for computers relying on visible pavement and lane markers, this is much harder.
While Google is still testing its self-driving cars on the snowless streets of California, Ford has announced in a press release a more ambitious project that will take on a bigger challenge. The driverless Ford Fusion Hybrid is using the LiDAR system to create high-resolution three dimensional digital maps of the infrastructure and road.
The four LiDAR scanners are mapping trees, buildings and road signs. The highly advanced technology is generating 2.8 million laser points per second and collecting up to 600 gigabytes of data per hour in the process.
The car is creating these 3D maps in ideal weather. For instance, it might create a map of the owner's commuting path during the summer. In the winter the car will use the map to locate itself on the road covered by snow. Ford claims that their technology allows locating the car with an accuracy of one centimeter.
According to the publication ECN, another concern with driverless cars in the snow is that the vehicle would get confused by the precipitation. Ford's high-tech sensors help avoiding this issue. They are designed to recognize rain and snow. The car would decide to drive through them, filtering them out of its "vision."
When ice, salt, snow or all three would block a sensor, Ford's LiDAR system would still be usable. The four sensors are combining all the data via sensor fusion and work as a team.