What challenges autonomous vehicles in the Arctic?

Heikki Konttaniemi (MBA) works as a team leader and project manager in Arctic Power RDI-group where he works with ITS projects. Hanna Kumpula (M.Sc. (Tech.), MA) works as a project engineer in Arctic Power RDI-group mainly with UI design and data visualization.

The Aurora Summit gathered 250 experts from over 20 countries to Olos, Finnish Lapland in January 2018. Such interest pinpointed that connected and automated transport has to take into account all conditions as it is not just for the good days. There were so many good speeches and presentations that it would be hard to summarize everything. Here are some thoughts about the summit.

Where’s my position?

What makes Lapland such a challenging place for testing autonomous vehicles and other transport automation solutions?

First of all, the global navigation satellite system (GNSS) is poor and unreliable. There are no GNSS satellites overhead in the Arctic.
For autonomous applications the positioning has to be even 10 centimeters and less. The atmospheric modelling for positioning purposes becomes hard due to the magnetic storms that we see as Aurora Borealis phenomenon.

Snow and ice are a problem

Another problem is the snow and ice. Snow and ice covers the road and the vehicle cannot use lane markings that could be otherwise detected and followed by using a camera system. To make things harder, especially during the winter, the road becomes narrower if there is packed snow or ice in the sides of the road, or ice that forms from the melting snow in spring time.

These conditions also often force people drive in the center. Would an autonomous vehicle avoid driving in the center and steer the vehicle to the slippery, uneven space on side of the road, if there’s another vehicle approaching? Then there are other issues with the so called “snow smoke” and drifting snow, as well as water in all its other forms, that easily makes cameras and lasers blind.

Go rural? Not so easy!

In rural areas such as Lapland, there are no HD maps, good telecommunication systems or even internet everywhere. These aforementioned things will not be prioritized either, since the population densities are low.

Still the autonomous vehicles need good connection – this is the focus actually in US where the government focuses on connected & autonomous vehicles (CAVs) instead of AVs (autonomous vehicles) or CVs (connected vehicles) alone. Car manufacturers such as Volvo need good 3G/4G/5G coverage, geofenced/digitized infrastructure, high definition GPS, local transportation control centre and a high defined map of the geofenced area also in rural cold climate areas.

We’re not alone with our conditions

However, we in Lapland are not alone with these problems. There are rural areas elsewhere, too, with similar concerns.

It is striking that for example in the US, 36 % of the roads are non-paved when looking at places such as Texas, California, Michigan, Nebraska and Alaska. These non-paved rural roads are very difficult to pre-drive and map while there are major objects and vegetation changing every season, not forgetting the varying road surfaces that are subject to different maintenance practices.

Therefore, Lapland is not the only place in the world that will have to wait for autonomous solutions in transport. Moreover, it was well summarized that Lapland acts as an issue amplifier – the challenging conditions that exist elsewhere are here, but amplified to another extent.

Steve Dellenbeck from SW Research Institute presenting some facts about automated driving in rural roads

Bad news: autonomous vehicles can ruin our roads

As if these challenges are not enough, the road deteriorates very fast if the autonomous vehicles take over and are programmed to have very little tyre wander. The amount of tyre wander has become less and less over time and will be potentially even less with autonomous vehicles that want to keep the exact same distance, for example to the road markings.

Imagine platoon driving of heavy trucks during a spring thaw. In such cases there might be even very severe pumping on the road and the costs of infrastructure maintenance would boom drastically. Our roads are not homogenous and this has to be taken into account when introducing autonomous driving and truck platoons in the AV-hype.

The “Road Doctor”, Timo Saarenketo explaining about the challenges of autonomous driving to road infrastructure.

The desk is not the right place to build insight

What to do? There are possibilities for us universities also, but as Dirk-Jan well said, “the desk is the most dangerous place to see the world from”. To come up with the right solutions, and to find our place, we need to build networks and relationships even more. There is plenty of work for everyone but they key is to work in an open environment and share data.

From a design perspective cases such as these should be designed to work in the worst environment first. Finnish Lapland is definitely an extremely good place to try and tackle different winter driving conditions.

In our opinion, as long as the autonomous vehicles cannot survive in every possible condition they can merely assist drivers. Furthermore, there will be different challenges in the transfering period and these were not much discussed at Aurora Summit. One of these is the safety idea where the autonomous car gives control back to the driver, if it cannot handle the conditions for any reason. The problem with this is that it takes about half a minute for the human driver to assess the situation and really take control safely.

This writing used the following peoples’ summit presentations for facts, thoughts and ideas:

Steve Dellenback, Vice-president of Intelligent Systems Division, SW Research Institute
Anders Eugensson, Director of Governmental Affairs of Volvo Corporation
Timo Saarenketo, the “Road Doctor”, CEO of Roadscanners Ltd.
Harri Santamala, CEO of Sensible4 Ltd.
Dirk-Jan de Brujin, Rijkwaterstraat, Director of Traffic Innovation Centre of Helmond
Jarkko Koskinen, Deputy Director General of FGI in National Land Survey of Finland