Ford is testing its fleet of fully autonomous cars in snowy conditions in a move that would attempt to conquer what is considered as one of the biggest barriers in creating self-driving vehicles.
In a recently released video, the company is seen running its Fusion fleet along the snow-covered streets in Michigan. In order to navigate the streets with accuracy, Ford said that it uses LiDAR data to gather major landmarks surrounding the car's vicinity, along with on-board maps built with high resolution technology.
"You know, our experience driving around...you know, we're here in Michigan. There's a lot of winter here. We're used to the conditions changing very rapidly. As humans, we understand that that can happen and we know how to deal with it. And we need to make sure that our autonomous vehicle can also deal with those situations in the same way," said Greg Stevens, Global Manager, Driver Assistance Research, Ford Motor Company.
Stevens added that one of the real strengths of autonomy is the sensing technology that is built all the way around the vehicle. After vigilant sensing is performed and data had been gathered, all of these are fed into the car's powerful computers which use a number of algorithms to come up with the so-called "what if planning."
"And then, when we have to do something like, say, making evasive maneuver, we can train the vehicle to do that maneuver. And so we can have the vehicle drive in those situations, especially slippery low-grip situations like there are in snow, in a way that an expert highly trained driver would have," said Stevens.
Ford is partnering with the University of Michigan's college of engineering for creating the 3D maps during dry weather which shall be used to keep the vehicles on the road. While these maps are not expected to exist at all times, it would seem like these autonomous cars could really take advantage of the technology when driving along roads it had driven before.
As long as the routes, particularly those which are covered with snow, have become familiar, driving on them should come more natural and much easier.
Google is also working on making its self-driving cars safer in bad weather, and the trend is expected to grow strong.
"We feel it's important to test in weather like snow. And not only to do the testing to understand how the system performs and how we can approve it but also to let the public know we are looking at these conditions and we're looking at how to make our systems robust to this kind of weather conditions," said Randy Visintainer, Director of Autonomous Vehicles and Controls, Ford Motor Company.
