In your doctoral thesis, you dealt with what is known as “platooning”. This means that fully autonomous and networked vehicles form a convoy (platoon) and drive in perfect coordination with each other. Platooning was originally developed for freeways to improve road capacity utilization, make traffic flow more smoothly and safely and reduce emissions. However, you have been working on networked convoy driving in city centers. What inspired you to do research in this area?
I attended lectures, seminars and wrote papers on vehicle networks, network simulations and mesh networks during my Master's degree. And through my jobs as a student assistant and professional positions after my Bachelor's and Master's, I always drove a lot - an hour's drive was the norm. If you drive through city centers yourself, it's very uncomfortable and annoying. So my choice of research focus was also influenced by the intrinsic motivation to potentially make my own life easier in terms of driving.
In what way? How can platooning in the city make life easier?
Platooning is originally an application for the freeway, and there is already a relatively large amount of research and field tests in this area: with platooning, the vehicles drive very close to each other, often around five meters apart at freeway speeds. This small distance between the vehicles means that the slipstream effect can be used, i.e. each vehicle drives in the slipstream of the vehicle in front. At the same time, the vehicle in front has no turbulent zone at the rear to pull it back. Trucks can save 10 to 15 percent fuel in this way, which is enormous for a logistics chain. In addition, the shorter distances mean that significantly more vehicles can be put on the road. In driving school, we learn that we should keep half a speedometer distance; at 130 km/h, that's around 65 meters, which is actually far too little in practical terms. However, if the distance between the vehicles in the platoon is only 5 meters, the throughput on the road increases enormously and traffic would flow much more smoothly thanks to the coordination between the vehicles. The so-called “phantom traffic jams” could be avoided or significantly reduced. We've all seen it: someone brakes hard, the car behind brakes too, and so it goes on. These traffic jams primarily occur because the distances are too small and people brake extremely hard. This problem can be solved by coordinated driving in a platoon.
And that is also the reason why platooning - at least according to my original idea - would also be very profitable in the city center. Here we have a similar problem at traffic lights. When the lights turn green and the vehicles start to accelerate, there are delays: The first car drives off, the second with a bit of a delay, the third again with a delay and so on. This creates these enormous gaps, which mean that the roads are not used effectively. However, if the cars were to drive in a platoon, i.e. in a convoy with only small gaps between them, the throughput at a junction could be increased, the journey time shortened and fuel consumption reduced because you would no longer have to stop at red lights so often.
A very promising idea. Has it been confirmed?
At the beginning, we first investigated whether the idea of using platooning in city centers would work at all. Then it turned out: If you apply the whole thing to a scenario that is typical today, i.e. a simple standard traffic light circuit where vehicles arrive, form a platoon and drive off, then you have a travel time reduction of just under 17 percent. Fuel consumption is also reduced by urban platooning. According to our analysis, this is also around 15 percent. This reduction is not primarily due to air resistance, but to the fact that the acceleration process may only take place two or three times instead of five or six - because the vehicles have to stop less often at traffic lights. This leads to less energy-intensive maneuvers.
With all these advantages, let's turn to the safety aspect: Assuming a mistake happens and the vehicles collide, this would be fatal at a distance of a few meters - compared to half the speedometer distance. In other words: If there's a crash in the platoon, it's a real crash, right?
That's right, today's emergency brake assistants cannot prevent this at such close distances. That's why wireless communication between the vehicles in the platoon is so important: platooning can only work if the vehicles communicate reliably with each other. And this is of course a greater challenge in the city than on the highway: on the highway, after all, you have an open field where the wireless exchange of information is often possible, as long as we rule out strong interference for the time being. It's different in the city. Here there are many obstacles that can interfere with the communication of the vehicles in the platoon, for example buildings, parked vehicles, trees and other objects. If, for example, a vehicle turns at a junction with buildings, the exchange of information between the vehicles is no longer reliable, the platoon becomes potentially unstable and collisions can occur.
That's why I investigated which innovative communication strategies can be used to reliably guarantee connected driving in a platoon, even in city centers.
How do vehicles communicate with each other on the highway? And to what extent do these communication methods reach their limits in urban platooning?
In highway platooning, there are procedures in which signals are modeled on the headlights. The light from the headlights then flickers extremely quickly - so quickly that it is no longer perceptible to the human eye. The same happens with the rear lights. So the first vehicle in the convoy with flashing rear lights can tell the following vehicle: “I'll brake in 100 milliseconds”. My first approach was to use this communication for the city center as well. It works very well - until the convoy turns off. Then the vehicle turning off no longer has perfect alignment with the following vehicle and information can be lost. A typical situation: I turn off, a cyclist crosses the road and I have to brake hard on the spur of the moment. If this information cannot be passed on to the rear in a platoon, collisions can occur. Of course, this must not happen. This shows that the approaches to communication for platooning on the highway are not necessarily directly applicable in urban areas.
Other standard procedures rely on an omnidirectional radio signal, i.e. transmission in all directions: This uses an antenna like the one you know from your router at home. The antenna “calls” information in all directions. In a platoon, however, it is often a matter of communicating to the rear and letting the next driver know what is happening. Today's standard procedures therefore also “call” in directions in which nobody is located. This is actually nonsense for platooning as an application.
Other methods are therefore required for urban platooning so that the vehicles can communicate reliably with each other. What innovative communication strategy did you research?
Our core idea was to bundle the energy and radiate it in the exact direction in which the communication partner is located. So we use a funnel that “sends” backwards. This works quite well at first, but it also has disadvantages, because a funnel like this is extremely loud. You can imagine it as if I were shouting at my conversation partner with a megaphone and the people around me would no longer understand the actual conversation. That has to be prevented. We experimented with how to position this funnel so that it doesn't disturb others. The bottom line of this experiment: the funnel must be positioned in such a way that it does not disturb other vehicles too much. Positioning it as low as possible, e.g. at the height of the front and rear lights, proved to be beneficial. It is also advantageous if one of these funnels is used for transmitting and a second for receiving. You then have a very sensitive hearing and a very loud call - and it turns out that communication with this system is not disturbed. You can imagine it as if the other person were shouting directly into my ear, and then I simply can't hear anything else.
In our simulations, we were able to enable perfect communication between the vehicles in this way, even when hundreds of vehicles were “interfering” in road traffic. This is particularly good for a safety-critical application. In various scenarios, we were actually able to receive 100 percent of the information that we sent. We have not yet found comparable results in other work.
This bundled transmission worked perfectly, especially when turning and in traffic jams. This was not possible with previous approaches.
Another question in your dissertation was what potential the urban infrastructure has for the coordination of platoons. What is your bottom line?
In my opinion, it makes sense to assume that infrastructure is not necessarily always available. A platoon should always function without any infrastructure at all. This means that the vehicles communicate directly with each other, e.g. via Wi-Fi or C-V2X (Cellular-Vehicle-to-Everything-Communication), which is part of 5G and also works without infrastructure. If I rely completely on infrastructure, platooning will no longer work if there is a failure, disruption or overload - for example at large public festivals, soccer matches or similar. A static infrastructure would also involve very high costs. However, it would be conceivable to use existing infrastructure opportunistically, i.e. if infrastructure is available, then I use it to take advantage of it. However, if it is not there, the system must still function sufficiently well.
Therefore, another research question in my work was to use the infrastructure of future smart cities opportunistically. The focus here was on the use of drones to relay wireless transmissions from vehicles. Our idea was to use drones that are already in the air. Platooning in the city center will not be possible overnight. If you then think further, it could be that delivery traffic is outsourced to the air - in other words, drones are used to deliver medicines, for example. If the drones are already flying up there anyway, they could perhaps also be used to relay wireless transmissions from vehicles and coordinate channel access across several platforms.
Can we look forward to urban platooning in the next few years?
There has been a lot of research into highway platooning for many years, but it is not yet ready for the road. Level 4 and Level 5 in particular, i.e. autonomous driving completely without any human influence, does not yet work reliably - even on freeways - and is not yet legally possible, at least in Germany. Until we have perfected the well-researched scenario on freeways, nothing is likely to change in urban traffic, which is still far too dangerous. A look at the accident statistics shows that 70 percent of accidents happen in urban areas and only 8 percent on highways. So there is a significantly higher safety risk if I allow autonomous or cooperative driving in city centers. So it will probably take a very long time.
Do you see potential for further research based on your dissertation?
In addition to the use of drones, satellites could also be used for vehicle communication. Low Earth Orbit satellites, LEO satellites for short, move at an altitude of between 160 kilometers and almost 2000 kilometers, delaying communication due to this distance, but are nevertheless an exciting building block for the entire field of vehicle communication and also for the coordination of several platoons.
Heterogeneous communication approaches are also exciting. For example, signal transmission via the front and rear lights in combination with a W-LAN approach. Depending on which technology is more suitable at the time, it would be possible to switch back and forth between them.
The human factor is also exciting. When it comes to user acceptance, relatively little work seems to have been done at present. Ultimately, even a technically sophisticated system is useless if nobody wants to use it in the end.
What are the next steps in your scientific career?
I am currently working in science transfer and talk a lot with companies, partners etc. about autonomous and connected driving. This has already resulted in many interesting contacts with whom possible joint projects have been discussed and concrete projects are also being planned.
Teleoperated driving is also an exciting area in which I have been active over the last two years and in which I would like to continue working, especially in a practical way. Connected driving or teleoperated driving in agriculture are also exciting areas. Here again, there are other special features than on the road and I can imagine many exciting and profitable applications in this area.
Title of the dissertation: “Cooperative Mobility in Urban Environments”, Commission: Prof. Dr.-Ing. Christoph Sommer (TU Dresden, previously University of Paderborn), Prof. Dr. Claudio Ettore Casetti (Politecnico di Torino), Prof. Dr. habil. Gerhard Weber (TU Dresden), Prof. Dr.-Ing. habil. Martin Wollschlaeger (TU Dresden), Jun.-Prof. Dr. Verena Klös (TU Dresden)
