Meet Eva the Self-Driving Bus: Auto-Vehicles Go Through Final Harsh Winter Conditions Test

This microbus even has a name - Eva. Its distinctive feature is that it doesn't need a driver. Eva is a self-driving car.

SPECIAL REPORT

ROBOT AT THE WHEEL, BY LEYLA ALNAZAROVA

This microbus even has a name — Eva. Its distinctive feature is that it doesn't need a driver. Eva is a self-driving car.

Andrey Vavilin, BaseTrack founder: "This is an innovative vehicle built on the chassis of GAZelle Next. Then, we automated it. We used our own approach. It isn't a classical approach at all. We don't have gauges and sensors, we don't use them as they aren't necessary. We generate very little data. A smartphone is enough to manage an automobile".

 

We didn't manage to talk to the operators of the self-driving GAZelle. We couldn't distract them as they were preparing for a test ride. Eva is one of 13 vehicles that takes part in the technological contest "Winter City". The challenges are carried out at the auto test track NAMI in Dmitrov near Moscow. The participants are to design a self-driving car that can drive along city roads in the conditions of Russian winter. They should be as good as an average driver. The main condition is that they shouldn't violate traffic rules.

It's more important to take part here than to win. The global goal of the contest is to create technological solutions with high investment prospects and export potential.

Kirill Kaem, Senior Vice President for Innovations of the Skolkovo Foundation: "In technological terms, mathematics, AI, and neural networks account for 80% of the success of those cases. Actually, Russian teams have always been ones of the best in the world."

A year ago, there were only three teams working on the creation of a self-driving car in the country. There are about 15 teams this year. The number has increased by five times over a year. But is it quick enough?

Alexander Povalko, RVC CEO: "We make decisions pretty slowly. I don't mean that there are long discussions, obstacles, and barriers. I compare us with our competitors. We should understand that there's fierce competition for the leading position in the development of self-driving vehicles".

According to analysts, legislative initiatives and investments are necessary for harmonious development.

Dmitry Peskov, Special Presidential Representative on Digital and Technological Development: "We can expect a serious economic effect within about 10 years. But in order for it to become a branch of the national economy within 10 years, we need to dramatically increase the sum of investments now. I'd say that we need more investments and more teams working on it by at least 10 times in five years. Government bodies and large corporations play a crucial role here. Then, after about 2025, there'll be a real application."

Assembling a car is only 10% of the work. It's important to gain experience of its usage in cities, on highways, and moreover, in Arctic conditions. A pilot project is already being implemented. They carried out experiments at the innovation centers in Kazan and Moscow. They plan to include other regions into the program up through 2022.

Denis Manturov, Minister of Industry and Trade: "Our main goal is to ensure testing of all of the components of self-driving cars on our country's roads in different climate zones and seasons within this period of time so that it would be possible to safely permit those cars to drive along public roads starting in 2022".

Self-driving taxis have been used in the Skolkovo innovation center for half a year. We can't call a taxi ourselves as one must be a resident to do so. That's why we'll ask a Skoltech student to help us.

- Hello, Nikita.

- Hello.

- Let's call a taxi?

- Let's do it.

- Hello.

- You need to buckle up.

- It won't drive without it?

- It must follow the traffic rules.

- Safety first! Okay, we're ready.

Nikita Veliev studies space technology in Skoltech. There are many buildings, and sometimes he has to take a marshrutka from one to another. The appearance of free taxis on the campus is a pleasant bonus. In addition, it's self-driving.

Nikita Veliev, student: "The main thing is that there isn't anybody at the wheel. It is a bit scary, especially in the beginning. You also feel your involvement in new technology".

The taxis operate in a test mode. The service was launched in August of 2018. Over this period, self-driving taxis have completed over 2,500 rides with passengers in Moscow and Kazan. But the company tests its cars not only in Russia.

Artyom Fokin, Director of Business Development for Unmanned Vehicles of Yandex: "We also have a permit to carry out tests in Israel. At the beginning of this year, in January, we attended the Consumer Electronics Show in Las Vegas where we also demonstrated how our technology can work on American roads".

Yandex's technology was highly appreciated abroad. The Russian company and the South Korean car parts company Hyundai Mobis will develop a hardware-software system. It'll be designed for self-driving cars (SAE Level 4 and 5). Level 4 cars are highly automated. Level 5 cars can drive from one place to another independently. A human just needs to activate the system and specify the destination.

The first prototypes of self-driving cars with a new hardware-software system can be designed this year. Yandex will probably decide to test them in its pilot project. This spring, they'll permit self-driving cars to drive on public roads. The Ministry of Industry and Trade has prepared a draft regulation. In order for self-driving cars to be permitted to drive on the road, a company-owner should have a charter capital of at least 100 million rubles. So, only big companies will be able to take part in the experiment.

Denis Manturov; "As for safety, we should think in these terms. What we suggest today will be considered at public hearings and be subject to traditional negotiation. I think that this document will be based on those proposals. Small firms can take part in the work on some elements. Today, the key leaders of the project are the NAMI, KAMAZ, and Yandex. We'll welcome other participants if they appear".

There's the important condition that a self-driving car should be controlled by an operator who will sit in the passenger's seat. All possible risks should be mitigated. However, Minister of Industry and Trade Denis Manturov doesn't rule out that there'll be even more revolutionary changes in the near future.

Denis Manturov: "I talked with the company Renault. I asked them how they handle the issues with the tests of their self-driving car. They offer bolder solutions. They're already using their cars on roads without escort by a human. I think that we'll probably introduce some changes or make more radical and bold decisions while carrying out tests".

In essence, a self-driving car is the same as a person at the wheel, only, it has an electronic brain. And it uses a set of equipment instead of eyes: cameras, radars, lidars, or sensors. Many modern production cars already have elements of self-driving systems. Take, for example, active cruise control, ABS, or active lane control. Some designers of robotic cars don't overlook the existing functions. They reach a new level by adding their own technology to it.

Andrey Vavilin: "We changed our approach to the development of self-driving cars. We need to come up with a safe driving solution with respect to objects, pedestrians, and other things. We have plenty of options, including the manufacturer's systems which are already installed".

The company's founder, Andrey Vavilin, has been specializing in geoinformatics for over eight years. His team specializes in modeling, cartography, geodesy, and navigation. The automation of vehicles was based on the expertise in those areas.

Andrey Vavilin: In technical terms, we prepare a road. The data should be set for a long time, it doesn't need to be updated. Then, from the office, we set virtual tracks along the roads. We call them basetracks.

- It'll drive the way you draw it, right?

- You're absolutely right. It drives the way we draw it. It goes along the virtual tracks that we create.

It turns out that the AI doesn't make decisions. It's explicitly told where and how it should drive. But will this approach be suitable for urban roads? Andrey Vavilin confesses that it won't. But that isn't their task.

Andrey Vavilin: "Our company develops in two areas. One of them is restricted areas, quarries, hazardous industry, routine routes where some movements repeatedly take place and there aren't any major changes such as objects, cars, and other things chaotically emerging. The other one is highways outside of cities. But we work on limited self-driving systems there".

Company's philosophy is simple. Technology should lead to the optimization of costs and higher profits.

Andrey Vavilin: "It should be scalable, simple, and cheap. And we actually make it now. Knowing the precise route, we can only take the accelerator from a driver and let him save on fuel and maintenance".

Such solutions can be used to transport cargo over long distances or in agriculture. But unmanned harvesting isn't a new idea at all.

Olga Uskova, President of Cognitive Technologies: "We started working in agriculture in February of 2019. As of now, we've signed agreements with five farms. I'm talking about Russia now as we started to cooperate with Latin America earlier. Starting in March or April, depending on the region, two or three robots will go to the fields to collect data and start working".

In addition, the Russian IT company is going to launch serial production of 4D-radars. This kind of device is being developed in Israel, Japan, and America, but Russia was the first to introduce this term.

Olga Uskova: "Of course, it sounds like a fashionable thing. In essence, it's a 3D-radar, but it shows a 3D-picture in dynamics. It recognizes an object and its shape when the object is quickly moving. It's either the radar or the object move quickly, or they both move quickly. That's why it's a 4D-model".

The company has already had an order from a car manufacturer. They don't disclose its name as it's a commercial secret. The first batch will be 300,000 radars. By 2020, they plan to produce up to two million of them.

Olga Uskova: "Actually, it's an even more significant Russian achievement than the launch of its own car, because if one launches his own car and a few people buy it... But I think that in this sense, new markets will be controlled by those who sell the brains, "nerves", sensors. I think that's the future. As a result, it's us who are in control".

Foreign countries already noticed Russian "brains". Korea, China, Germany, and Italy are making pre-orders.

Yury Minkin, a department head at Cognitive Technologies: "The key difference between our radar and a classic one is that our radar shows 3D-picture and time. We can see an object, its shape, its size. We can understand what we see — a truck or a bridge. Based on that, we can make better decisions, classify an object more precisely, know what object it is and what characteristics it has".

If you look at the internals of a 4D-radar, there are components made abroad such as chips. But many of the components were made in Russia.

Olga Uskova: "All of the automatics were made in Russia. They were made in Moscow, Tomsk, by teams from Saint Petersburg. We can say that so many smart people worked on it that it consists of about 50 years of human knowledge".

Developers use cameras, radars, or lidars, or combinations of them, or think up of something and walk away from sensors and gauges, using unconventional approaches, trying to create a perfect robot at the wheel which will be able to ensure the most important thing, safe driving. Despite all of the progress made, we're pretty far from the widespread usage of self-driving cars. The appearance of them on all of the roads in the country is real and probably inevitable. But it's more likely a picture of the future which is now being worked on.