Road traffic is facing a radical transformation: the future belongs to a system of autonomous vehicles connected with artificial intelligence. This is intended to solve traffic problems. But what are the implications for cycling as a means of transport? Will the disruptive influence of humans be excluded in the future?

Ninety per cent of all road accidents are a result of human error. If the human factor is minimized, mobility will become safer – this is one main argument presented by the supporters of robot cars, which are currently not only undergoing initial test drives in US cities but also in Europe. Theoretically, a connected, computer-based system will also provide more safety for cyclists. Yet in reality the picture has so far been rather different. During a test in San Francisco, self-driving vehicles turning off at a junction crossed a dedicated bicycle lane without paying attention to the cyclists. In road traffic it is also usual for cyclists to communicate with the driver of another vehicle by eye contact or indicate a change of direction by hand signal. Both of these are difficult processes for sensor-controlled systems to comprehend. They need clearly defined traffic situations. This is why traffic planners hold out little chance in future for play and cycle streets as well as shared space areas, as they prove to be overtaxing for autonomous vehicles with the current state of the art. Consequently, it would be necessary to separate the individual types of traffic. Putting an end to the cycle as a means of transport as we know it today also seems to a conceivable option for them.

Autonomous cars have no problem with cyclists. On the contrary, if cyclists behave sensibly, autonomous traffic is even safer for them

“Autonomous cars have no problem with cyclists. On the contrary, if cyclists behave sensibly, autonomous traffic is even safer for them”, reassures Armin Müller, Managing Director of the company EMM! Solutions. The company is working on its own autonomous vehicles and is also dealing with the interaction of vehicles in public space, e.g. on a test circuit at the Siemens Campus in Munich. One point to note is that autonomous vehicles drive significantly more passively because they behave in exactly the way they have been programmed. In contrast to human drivers, they observe all traffic regulations and have no moral sense of being disadvantaged. “The greatest problem will be the difference in speed. When an autonomous vehicle meets a slow cyclist, the first step is to adapt to the speed of the slow road user. That certainly won’t please every occupant”, explains Müller.

In spring of 2018, an autonomous Uber taxi ran over a while it was dark. Changing light conditions or external influences such as fog, in particular, are still difficult to judge. Many test drives take place during good weather conditions. However, the known accidents stir up fears about technical progress.

New research projects are aimed at gaining a better knowledge of cyclists’ behaviour and dealing with special cases. While overtaking cyclists with a sufficient safety margin now happens as a matter of course if the cars have been correctly programmed, initial trials now show that autonomous vehicles can even adjust to cyclists weaving in and out. They are even overtaken with a greater safety margin than is legally prescribed. And the bicycle does not even have to be connected within the system – although that would certainly not be disadvantageous. “A lot of opportunities only materialize once all vehicles are communicating with each other. Connectivity will also become increasingly important in the bicycle sector and will require constant and ongoing development”, is the conviction of Sven Bernhardt from Comodule. The company from Estonia presented an autonomously travelling app-controlled e-bike back in 2015. While the project has not been pursued further, it is by no means some crazy vision of the future.

After all, researchers at the University of Magdeburg are currently examining a self-driving e-cargo tricycle. After undergoing trials on a test circuit since early 2018, the cycle is now ready to commence the next step on 1 July 2019. Over a period of three years, the test area is to be expanded to include the campus and surrounding streets. In this process, the cycle first has to learn how to communicate with its environment. “That begins as soon as it has to face an oncoming cyclist. The trike has to consider what will happen and how the oncoming traffic might behave”, explains Stephan Schmidt, who is jointly responsible for the project. In order to be perceived itself as well, the trike also has to learn how to make itself noticeable, such as when it changes direction or when pedestrians are blocking the path. “The tricycle is spatially closer to the pedestrian than autonomous cars. The questions of perceiving the environment are even more relevant”, explains Schmidt. Before the project could be implemented at all, various maps had to be created for the cycle path infrastructure. In contrast to the situation for car traffic, there is no comprehensive network and, instead, a confusing hotchpotch. In the case of autonomous e-bikes, an insufficient cycle infrastructure remains a source of conflict between cars, bikes and pedestrians.

Schmidt envisages the bikes as primarily being used for local public transport. Commuters travelling by bus or urban rail can organize the appropriate e-bike for the last mile of their journey, which is then ready and waiting at the stop and shortens users’ journey times. In contrast to autonomous car traffic, this could significantly reduce the burden on cities. After all, although autonomous cars travel more closely behind each other, can maintain the lane and speed better and park together more tightly than human drivers, first computer studies show that the traffic will not be reduced – but will instead increase. Tesla CEO Elon Musk said that Tesla drivers would in future be able to earn 30,000 dollars per year because their autonomous car would drive for Uber when not in use. The resulting increase in empty trips would be an additional burden on the infrastructure. In return, though, parking space would become superfluous, which would benefit cyclists, pedestrians or even parks and green spaces – but only if the municipalities were prepared to provide the funding.

Autonomous vehicles will completely change the type and manner of traffic in future. Precisely what form this will take is still difficult to forecast today. Traffic expert Chris Gerdes from Stanford University speaking on the Deutschlandfunk radio station observed: “Today the road belongs to cars, and pedestrians have to give way to them. In future that could be completely reversed.” It is therefore also important for pedestrians and cyclists to communicate with cars. They could then use an app to request a zebra crossing where they want to cross the road. According to Gerdes, there would be many different options for a very different way of organizing traffic.