Considering the fact that the company has announced it is planning to release at least semi-autonomous vehicles by the year 2020, Ford is heavy at work with testing the driverless car systems to make sure they are prepared and flawless by the time launch is imminent. While most driverless car testing is normally done on sunny, clear-skied days on empty highways in California, to ensure the most accurate results.
But unlike Google, Audi and Mercedes who do choose a nice, clear day before they take the cars of our future out for a driverless spin, Ford doesn’t seem to be phased by the harsher conditions of Mother Nature. Just recently, Ford did various tests to determine how their autonomous vehicles cope with wet and slippery roads.
Truth be told, this is an issue that not many other have attempted to touch down on so far. But in reality, once these cars become a thing of the present and you and I and our grandparents will own one, it will surely not be all sun and joy on a daily basis and for the rest of eternity. Not to mention anyone living in colder regions that record higher precipitation rates.
Thus, Ford developed their autonomous car’s systems to make use of the same Lidar technology implemented in most planned driverless cars. Lidar is a system that makes use of lasers in order to map the environment in a 3D plane in order to let the host know exactly what are the surrounding objects, as well as how far away from it they are. Based on this type of information, autonomous cars can detect obstacles and know when and how to avoid them.
However, thanks to Ford’s testing of the system in bad weather conditions, researchers have come to realise that the Lidar sensors were functioning too well; to elaborate, the lasers actually ended up picking up snowflakes and raindrops as obstacles. This, in turn, made the car believe that precipitation should be avoided. That ought to have been horribly confusing for the car.
A partnership with researchers from the University of Michigan will hopefully result in an algorithm that can detect precipitation as being the kind of information that requires filtering out from the Lidar sensor feedback. That way, the car would simply be able to tell the difference and cease trying to avoid rain and snow.
Image Source: 1