Swedish company pushes autonomous cars to their limits to ensure trust in technology

The race to get autonomous vehicles ready for the road is creating new opportunities for companies. AstaZero CEO Peter Janevik says his company is operating one of the world’s most advanced test facilities for self-driving technology. He explained what makes the Swedish test track unique and outlined why it’s so important to be transparent about what autonomous cars can and can’t do during an interview with Automotive News Europe Correspondent Olive Keogh.

Why is your advanced test facility for autonomous vehicle technology in Hallered, Sweden, different?

It is the only track of its kind in Europe and the most advanced in the world. We are setting a new benchmark in automated vehicle and active safety testing by creating a unique mixed-reality test environment that really challenges automated vehicles and self-driving cars.

Does the track look like a normal proving ground?

No, because we have none of the things you would expect such as ride or handling tracks. Instead, we have a small city environment with an intersection and some scenery from Harlem in New York actually. But, we can simulate virtually any traffic situation in any city -- the Arc de Triomphe traffic system in Paris for example -- and blend it with real vehicles and real people positioned around the track. Being able to do this means we can create a completely new level of traffic complexity to look at how autonomous vehicles will share roads with other vehicles, cyclists and pedestrians.

Who owns the track and who are AstaZero’s customers?

It is owned by RISE (Research Institutes of Sweden) and Chalmers University of Technology. Our customers are car manufacturers and Tier 1 suppliers in Europe and other parts of the world.

AstaZero is the first company to secure permission to test across the full 5G frequency. Why is this significant?

If you have connectivity between cars and between cars and infrastructure and you have several vehicles coming up to an intersection, for example, you can negotiate who should go first. You can do away with traffic signals altogether and massively increase the throughput at the intersection. Connectivity is central to this and 5G is one of the key components that comes into play when vehicles and machinery are communicating wirelessly because it provides a substantial boost to data capacity.

Will self-driving vehicles have an overall effect on society?

Yes, because the attributes that are important today such as ride and handling will no longer be relevant and other things will come into play such as trust and behavior. This means there is a need for research not only into the pure technical aspect of self-driving vehicles, but in multi-disciplinary areas including behavioral science and how do we build trust in an environment with multiple vehicles all controlled automatically. At AstaZero, behavioral scientists can do research on equal terms with traditional vehicle experts. Both branches are needed to develop a solution that is really good.

Is trust a big issue in this sector?

Yes and we need to proceed with caution and to be very transparent about the capabilities of self-driving cars and what they can and can’t do. We need research into all the elements that go into building trust because trust is something we take for granted with human drivers.

What are the challenges that remain for self-driving cars?

From a testing perspective I would say robustness of sensors is one example. We need them to be reliable in Irish rain or Swedish snow or in complicated lighting conditions in Spain. We need to develop repeatable testing methods to simulate this on the proving ground and in the labs. There are a long list of challenges like this we need to solve. A car without a steering wheel that can handle any environmental conditions is still a long way away. There is also a need to standardize test control protocols so that tests in Sweden can be compared with tests made in the U.S. or South Korea. We need to arrive at a point where different sensors will all recognize and agree that the “target” ahead is a car.