Autonomous snow ploughs: The next test for driverless technology
Technology and Business
In this article, we consider the limitations of adverse weather on AVs and ask:
• Are autonomous snow ploughs in operation right now?
• How do driverless vehicles work in the snow?
• How might AVs prove useful in adverse weather?
Weather has long formed a part of the debate over autonomous vehicles (AVs) as their reliance on various cameras and sensors to understand the world around them naturally means safe travel in dry, light conditions is easier to achieve. Conditions like fog, rain and snow can make it impossible for driverless cars to detect lane dividers and road signs and could even obscure their own cameras. Though more data is being collected in a range of conditions, there’s no denying that fair weather is a default for testing in the industry.
Where driverless cars have been put to the test in snow, weaknesses have been discovered. For example, when Krysztof Czarnecki tested a self-driving car in Canada using data gathered in Germany, he found he had to correct its movements several times owing to the volume of snow on the road – something that just wasn’t present in the data. Make no mistake, humans are worse drivers in these conditions too – accidents actually increase by 20% in the winter months. But public desire to see AVs as a safer alternative to human drivers means they need to do better, not merely equal our efforts to navigate through adverse weather.
As Alexandr Wang, CEO of Scale AI, which regularly works with autonomous driving companies, says, “It’s a very noticeable blind spot. Deploying autonomous vehicles in bad conditions is not really tackled, or talked about.” Yet, for some vehicles, heavy snow could be considered a ‘normal’ driving condition and even here autonomous technology is making inroads.
Are autonomous snow ploughs in operation right now?
Snow ploughs are perhaps the ultimate bad weather vehicles, designed to work in the most difficult conditions, and already autonomous versions are being tested. Designed to work as part of a fleet which could clear deep snow in record time, Daimler’s first autonomous snow plough was tested back in 2018 to great success. Yeti Snow Technology soon followed with a system aimed at clearing snow from airport runways and highways. Tested for the first time at a snowy airport north of Oslo, the snow ploughs are impressive vehicles – 20 metres long, 5.5 metres wide and able to clear an area of 357,500 square metres (approx. 50 football fields) in just one hour.
Their ability to tackle tricky conditions all comes down to design and forward-thinking. Snow, and heavy snow at that, is accounted for from the very earliest stages of design in these vehicles – a consideration that is perhaps not afforded to other autonomous vehicles we have seen thus far. Yet these large-scale snow ploughs use much of the same technology as the zippy cars self-driving around places like California.
You want to know more about how this data could improve autonomous cars and their ability to do a better job at driving in the darkness than humans? Read our article "Prince of Darkness: Autonomous vehicles at night".
How do driverless vehicles work in the snow?
GPS perhaps becomes more important in these designs as, despite all efforts, snow still has the ability to block visual cues for cameras and even make sensors less accurate. The Yetis, for example, follow pre-programmed routes to minimise the potential of being guided off course. Though these routes can be adapted and the vehicles are connected to remote human safety drivers, the lack of additional cameras and sensors means they have less autonomy than other AVs.
However, other snow-clearing vehicles already include a wider range of sensors. In the US, Winnipeg Richardson International Airport is now home to an autonomous snow plough, nicknamed Otto, which uses a combination of GPS and LIDAR. Shawn Schaerer of Northstar Robotics explains, “We use GPS to guide the vehicle up and down the paths we want to maintain and we use sensors like LIDAR and radar to sense around it. If there’s something in front of it, it stops.”
This is an interesting point as, due to the very nature of the work these vehicles are designed to carry out, if weather conditions become too difficult to navigate, they can simply stop. An autonomous car driving on potentially busy roads, however, cannot. This is exactly why weather testing for autonomous vehicles remains such a hot issue as the fail safes (i.e. stopping) which protect autonomous snow ploughs, do not work for more commercial forms of transport.
However, testing of AVs is beginning to consider weather more prominently. Where most testing was once carried out in bright, dry conditions, companies like Ford are now going out of their way to ensure driverless cars see rain, snow, fog and greyer skies on test runs. There is a way to go but making weather a core consideration of testing and development is certainly a step in the right direction.
How might AVs prove useful in adverse weather?
Operating in all weather conditions may be a challenge for autonomous tech and the fact these snow ploughs are designed to stop when conditions become too dangerous suggests that the formula is yet to be perfected. To see the rollout of autonomous features in vehicles which must carry on through all weathers, significant testing and the development of sensor systems which can do their job in even the roughest weather is needed. But the success of driverless snow ploughs shows we are moving forward.
Mikael Carlsson, former CEO of Swedish logistics firm Kallebäck AB, sees that there is already a place for automation in haulage across colder climates. He explains, “Winters tend to be long and harsh in Scandinavia and this results in many traffic challenges. Often in extreme weather, technology is the first thing to fail. Sweden has many stretches of road with poor lighting, few road markings, black ice and substantial ice drifts.” Automation, executed well, could help to overcome these challenges, but based on current evidence, he’s not sure it’s there just yet, saying “It takes an experienced driver to navigate these and at the moment, I’m not convinced the tech is up to the challenge.”
Whether it is ready now or still a way off, transferring driverless technology into snowy conditions will undoubtedly bring benefits to all. The ability to clear airport runways will mean freedom of movement which is currently often blocked by adverse weather. It could mean more efficient delivery of goods, safer journeys on otherwise dangerous roads and will further open up travel in places where snow is a common sight. We’re excited to see where the road might take us.
Do you think snow-clearing is a good job for AVs? Would you trust a driverless vehicle in poor weather conditions? What could safe travel through snow mean for you? We’d love to know what you think. Share your thoughts in the comments below.
Submit your story
Become part of our autonomous revolution and submit your stories, images and videos
Stay up to speed with our weekly briefing. Enjoy autonomous driving content direct to your inbox