Driverless buses are the future.


Johannes Liebermann is the manager of “auto.Bus – Seestadt“, a research project funded by the BMK (the Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology). He sees quantum leaps in the area of self-driving busses.

For more than a year, Wiener Linien has been operating two small, self-driving electric busses at Seestadt Wien. The unique test conditions allow them to gain experience in autonomous driving in city traffic.

The project titled “auto.Bus – Seestadt” is run by Johannes Liebermann, who studied ITS (Intelligent Transport Systems) at FH Technikum Wien. At AustriaTech, a subsidiary agency of the ministry of traffic, he first established an office for automated and networked driving which serves as a national platform. The two Navya bus models give Seestadt residents a first glimpse into the “mobility of the future”. Normally, up to ten passengers are allowed per bus. Due to the current Covid-19 situation, the number of passengers is reduced to a maximum of five people from the same household or two passengers from two different households. An operator is present to supervise safety to traffic. Apart from Wiener Linien and Navy, the project also involves the Austrian Institute of Technology (AIT), the Kuratorium für Verkehrssicherheit (KFV, Traffic Safety Board), TÜV Austria and Siemens Mobility.

Why is Seestadt Aspern especially suited for this research project?
Johannes Liebermann: For this project, we were looking for an environment where public transportation is already effectively organized, but the two self-driving busses would still offer an additional service to residents. The busses run as emission-free, battery-powered vehicles. Due to the enormous computation, the route cannot be longer than three kilometers because all data must be processed in real-time. With all these requirements, Seestadt offers the best conditions because it is an urban environment which is still less complex than the routes at the city center.

What is the most fascinating part of the project?
Johannes Liebermann: It is the first time that our passengers can experience a close-to-prooduction and highly automated vehicle. Up to now, all advanced attempts of autonomous driving have been limited to closed systems with manageable complexity – such as the motorways, subways etc. A self-driving bus in public city traffic surpasses other automated driving projects.

The project is supposed to develop sensor technology, IT security and safety to traffic, among other things. What do you focus on?
Johannes Liebermann: AIT is our scientific partner in this project. Their focus is the vehicle’s added functionality. Currently, the bus feels its way, so to speak, via GPS and 3D LiDAR sensors. For the bus, every obstacle is like a concrete wall, even if it’s just a piece of paper flying through the air – the bus always plays it safe and stops. In the long run, an integrated 3D camera system will help the self-driving bus to recognize the type of obstacle and therefore better estimate its danger.

Is that the reason why it is necessary for an operator to be on board?
Johannes Liebermann: Whenever the bus stops, the operator checks if there is any real danger or if the bus can go on. But the operator also needs to be present for legal reasons. The system is driving autonomously, but there are numerous predefined safety steps such as stop signs or yield signs where the vehicle can only continue after confirmation by the operator.

Which insights did the project give so far?
Johannes Liebermann: The camera system is supposed to help the artificial intelligence to learn how to recognize different obstacles. But the example of the piece of paper illustrates how complicated this really is, because no piece of paper is identical to the next. We calibrate the cameras so that they are always positioned identically, accurate to a nanometer. We are currently evaluating the field tests to find out what the computer learned. When a small child learns something, you can find out its level of knowledge by asking. The structures of artificial intelligence learning are much more complicated because they are not as clearly traceable or retrievable.

When could autonomous driving in public short-distance traffic be used in all of Vienna?
Johannes Liebermann: There are experts who think that self-driving cars could be on the road error-free in the year 2040. However, this only applies to the motorways, because this environment is much simpler. All cars drive in the same direction at homogenous speeds. In the city, the situation is more complex. There is opposing traffic, different types of road users, pedestrians who could cross the street any time, and so on. It will take a lot of years until autonomous driving will be used in public short-distance traffic.

Does autonomous public transport driving also work in bad weather conditions and at night?
Johannes Liebermann: Darkness is no problem because it makes no difference for the GPS and the laser scanners. Bad weather conditions are a different matter. Light rain can be managed, but the bigger the raindrops are, the sooner we return to the “concrete wall” problem, and the bus cannot distinguish a drop from a big obstacle. There are “diffusion” research approaches, but this technology is extremely expensive.

Apart from that, where do you see further challenges for autonomous driving in the city?
Johannes Liebermann: Organizing stable computing power is a huge challenge. There currently isn’t even a technical environment that runs 100% stable. Think about how often you need to reboot your smartphone. On a route from A to B in a city, there are many unknowns and variables which could crash a program. Even somebody opening a window in a house next to the road and the reflection glaring into the sensor might cause a problem. Compared to the complexity of city traffic, running an automated nuclear power plant is a piece of cake.

The project is acclaimed in Austria. What is the international response? Does the rest of the world take notice of auto.Bus?
Johannes Liebermann: The research project is very closely followed internationally. Austria is a member of VDV, the Association of German Transport Companies, where we continuously exchange information with other shuttle projects. The model we are testing is Navya’s fourth vehicle generation, and we contributed a lot to its development. There are many delegations, especially from Germany and Eastern Europe, who come to Seestadt, and they are enthusiastic about our situation. For Europeans, Seestadt is a paradise for technology developers because they do not only find the infrastructure, but also good test conditions. It is an urban lab. Our project also creates great responses because the shuttle busses actually meet on their route, and you can watch them react to each other.