Experts: Carolin Holland (SBB), Martin Neubauer (SAAM)
Autonomous vehicles have the potential to become a game-changer in mobility. Firstly, however, they must overcome technical and regulatory hurdles. Secondly, they can only have a positive impact on the environment and help achieve sustainability goals, is if they are used in a predominantly collective manner.
Picture: Bernd Dittrich, Unsplash
Highly automated vehicles (SAE level 4) are vehicles that can completely take over the task of driving on certain routes, under suitable light and weather conditions or in specific traffic situations: within so-called ‘operational design domains’. If such a section of the route ends or an unforeseen event occurs, i.e. a fault, the car autonomously brings itself in to a state of minimal risk, brakes or parks itself. The driving must then be taken over by the person in the driving seat. Autonomous or self-driving vehicles (SAE level 5) can completely take over the task of driving in all situations and, as current design studies show, sometimes no longer even have a steering wheel.
As highly and fully automated vehicles are data-processing machines, it is not surprising that their development is mainly being spurred on by large software manufacturers, despite the conventional car industry having caught up in recent years.
In terms of technological development, the origins of autonomous vehicles are not to be found in assistance systems, even though these are indeed used, but in research on autonomous robots: For a conventional car to be turned into an autonomous one (actually a large autonomous robot) it only needs to be equipped with enough computers, data-analysis methods, navigation aids, sensors and actuators. However, it has recently become clear that these components represent a greater challenge than previously thought. Given the many possible scenarios and situations in road traffic, the process of automating vehicles is proving to be a very complex undertaking. Accordingly, there has been a certain amount of disillusionment in recent years, not only in public discourse but also among experts, regarding the vehicles’ launch timeline.
There are not yet any autonomous road vehicles of SAE level 5 on the global market. The first manufacturers are advertising their mastery of SAE level 4. These include the major American manufacturers Cruise and Waymo, as well as the Israeli Intel subsidiary Mobileye. In a pilot project conducted by Mobileye and Deutsche Bahn, 30 to 40 highly automated vehicles are to be used in and around Frankfurt. After initial trials in Switzerland, a new research project is set to start in Schaffhausen. This project is to be jointly run by the Finnish company Sensible 4, Toyota, Swiss Transit Lab and the Swiss Association for Autonomous Mobility (SAAM). It involves shuttles for local public transport that drive autonomously on predefined routes, in other words SAE level 4 vehicles.
Provided they are linked to public transport, autonomous vehicles could also become a mobility game-changer in Switzerland by enabling new and innovative mobility concepts. If autonomous vehicles are used collectively, rather than on an individual basis, they could save space and increase efficiency: in parking space, in capacity gains on the road, and in management of the overall traffic volume. One promise associated with autonomous vehicles is that of increased road safety via reduced accidents. It is not clear whether this can be realised. If it were to be realised though, transport’s cost to society would decrease.
In the USA, several highly automated vehicles are already in commercial operation. In Switzerland, this is not yet the case, due to the legal situation. However, the first test fleets have already been set up.
It is currently becoming apparent that business models are certainly possible with automated vehicles that are controlled remotely, i.e. by means of teleoperation. In particular, there are advantages in very specific areas of application: e.g. in factories, in warehouse logistics or in private environments, for instance with automated valet parking. These business models are perhaps not as groundbreaking as the manufacturing corporations’ marketing departments promise.
With regard to the development of self-driving vehicles, there are major challenges that cannot be described as insignificant, for example in the early and reliable detection of objects, especially in poor weather conditions. Rain, fog or snowfall can severely disrupt the sensors. Data processing has to take place in real time, which is no trifling matter, given the amount of data generated: BMW puts it at 40 terabytes per car per day. Due to the large number of possible situations, the automated driving system must also be able to react to situations that it has never been confronted with, even in simulations, and for which there is hardly any data. Another challenge, which affects the infrastructure in particular, is that of providing fail-safe internet connections a sufficiently high data capacity over a wide area (see article 5G applications). Furthermore, there are big question marks over introducing such vehicles into road traffic, for instance when it comes to proving their functional safety and roadworthiness. With regard to the organisation of road space, there is a need to adjust legislation. How human-powered transport, non-automated vehicles and highly or fully automated vehicles will coexist is a question that has yet to be resolved and is the subject of many studies. It is also still unclear how autonomous driving will affect overall mobility. Considering the advantages of autonomous vehicles, it is conceivable that the cars will cover a greater distance per year, which in turn will have a negative impact on the environment.
For projects in the field of autonomous vehicles, Swiss funding is at a very low level compared to Germany and the rest of Europe, and is only sufficient for basic research, not pilot projects. Moreover, this is while Swiss organisations are still benefiting from EU funding.
However, alongside financial support, what is needed most of all is clarity in the regulations governing the conditions under which the technology can be rolled out in a controlled manner. This is partly resolved by the revision of the Road Traffic Act and the corresponding ordinance. It would also be desirable to see increased cooperation between the cantons and pilot projects, so as to prepare society for the change.