Adverse weather conditions have long been considered one of the biggest barriers in the development of self-driving vehicles. Now, Ford has announced that it’s been testing its autonomous cars in the snow.
Most autonomous cars use lidar to generate an understanding of the road conditions around them, but snow makes that far, far harder to achieve with accuracy. Instead, Ford is using the lidar data to pick out major landmarks in the vicinity of the car, then using high-resolution on-board maps to navigate the street.
Those prior maps won’t always exist of course, so a technique like this would only really work on roads that the car had driven before. That’s something humans will be able to understand all too well, though: driving in thick snow is far easier when you already know the route. [Ford via Engadget]