SD Times Blog: An all-weather, autonomous car

by Jeremy

Please leave it to Finland to lead the way for self-driving car technology. While the company has no pure native auto manufacturers, it does have a thriving B2B market for parts and systems sold to automakers in other countries. We have yet to develop reliable, fully functional self-driving cars here in America, though we are getting close. Features such as blind-spot detection, 360-degree cameras, stay-in-lane assist, and automatic braking in cruise control mode tell us we’re getting closer.  But one thing we haven’t been able to overcome for fully autonomous driving is what happens when rain, fog, ice, and snow “blind” the sensors and cameras. The Finnish self-driving technology company Sensible 4 was hoping to learn as it develops its self-driving technology called Dawn. 

Screen Shot 2021 02 12 at 12.21.50 PM

Sensible 4 took first place in the Dubai World Challenge for Self-Driving Transport in 2019 and the Finnish Engineering Prize in 2020. With $7 million in funding from Japanese investors, the company expects to have its self-driving vehicles on the road next year. To make this happen, the last December took a car to Finnish Lapland to test it for two and a half weeks, in temperatures below -20 degrees Celsius (-4 degrees Fahrenheit). That’s way cold. The weather, the company reported, mainly was dark, snowy, and mean, with snow covering driving lanes and their surroundings and visibility dropping to a mere few yards at times. 

With this testing, Sensible 4 wanted to see how reliable the hardware would remain in arctic weather, test the entire software stack, and train data gathering. The company said that these whiteout conditions, where lanes, trees, and parked cars are covered in snow — along with people looking different in heavy winter clothing — were “an important aspect for the development” of the self-driving car.

Antti Hietanen, a senior autonomous vehicle engineer at Sensible 4, wrote in a blog on the testing: “For complete stack testing, we had designed common scenarios from traffic, such as vehicle overtaking, emergency braking, and adapting to a lead car, speed. In [Lapland], we were especially interested in conducting the test while the road surface was slippery and during heavy snowfall. Both are essential factors as during a wheel slippage, the vehicle motion calcula, ted based on the wheel encoders does not match an absolute vehicle motion a night confuse our vehicle localization. In addition to slippery surfaces, a heavy snowfall will pair our vision-based localization and object detection by decreasing visibility and deforming the landscape indistinguishable.

Related Posts

Leave a Comment