What if we used cars to detect other road users and predict traffic? The “IoT in mobility” project explored society’s need, the possibilities and the possible barriers.
Modern cars are increasingly sophisticated and packed with sensors, even the cheaper car brands. This use case focused on using cars to detect cyclists. Cleverly combining information about where cyclists are riding, how fast and in what direction can make intersections safer. Sascha Hoogendoorn, director of the Mobility Innovation Centre Delft: “In the Netherlands we have an enormous number of cyclists. Here in Delft, we witness daily that intersections could be safer. That helped us properly map different scenarios. We focused on combining data from different sources. After all, what is the power of 5G and how do you combine all the available data to get a good and reliable picture of the traffic?”
Combining and predicting
Cars are becoming increasingly capable of mapping their surroundings. For example, they can differentiate between cyclists, scooters, cars and trucks. They also register all kinds of other readings, such as temperature and road surface quality. The idea behind this research, is that multiple cars can spot all cyclists in their immediate vicinity. Thus, information is exchanged from various directions about where cyclists are riding, how fast, in which direction and the distance to the traffic light. If you match this with information from detection loops under a bike lane, you can make sure, for example, that a traffic light will remain on green longer in case of high crowds.
Combining the data and predicting when cyclists will enter an intersection offers all sorts of advantages. But it also brings dilemmas. Sascha: “We want to prevent accidents as much as possible, of course. So the light stays green a little longer, to avoid hasty cyclist taking a risk by running a red light anyway. It’s annoying for cyclists to stop and dismount. We have calculated that cyclists in an urban environment use 63% less energy if they do not have to stop at traffic lights. Besides safety, sustainability also factors in: we want to keep traffic, overall, flowing as smoothly as possible. On a large scale, designing systems more efficiently saves a lot of stopping and starting moments. But you have to make decisions about who receives priority at an intersection. At MICD, we spend a lot of time researching these aspects. For this particular use case, focused on cyclists, we mainly looked at what data we could collect and how to process it.”
Municipalities are eager to create attractive, greener inner cities, reduce car use and encourage healthy behavior. That implies giving more priority to cyclists and pedestrians. A consequence of giving cyclists and pedestrians longer time to cross, means longer red lights for cars. Sascha: “That means that cars will be waiting longer. However, we can’t keep them waiting too long. That would result in too long queues at traffic lights and angry motorists. Those long line-ups are also bad for sustainability reasons. So we looked at how you can cleverly favor cyclists and pedestrians, without putting car traffic at too great a disadvantage. If you have a better understanding of when a cyclist – or better yet a group of cyclists – arrives at a traffic light, you can adjust the traffic light accordingly. The largest challenge will be to find the proper balance between who gets priority and which group has to wait. Combining data analysis and knowledge about human behavior is, in my view, essential to designing the cities of the future.”
Power of 5G
Creating an up-to-date and reliable picture of all traffic participants is essential to make the right prediction and decision on what the traffic lights at an intersection should do. The deployment of 5G will be essential in the future because of its low latency, combined with the processing of huge amounts of data, to quickly calculate super fast where each cyclist is located.
Developing and testing an algorithm was one of the main objectives. This is an important part of ensuring that intersections become safer and cyclists can travel from A to B more smoothly. Yet there is another major challenge: the market is still quite conservative. “Suppliers of traffic control systems are still reluctant. They currently see no need to change their products yet. Understandable, because it will require some investment; not only financially, but also in time and energy. If we want to make traffic safer and more sustainable in the future, it’s not just about developing smarter technology. Applicability is essential. Suppliers must be able to understand the added value in order to be convinced of usefulness and necessity,” Sascha points out.
Intelligent Traffic Light Controller(iTLC) installations, also known as intelligente verkeersregelinstallaties (iVRI) in Dutch, are control systems that facilitate communication between the traffic infrastructure and road users, allowing for the collection of more data than previously possible. This enhanced data collection capability leads to improved and innovative traffic management possibilities.
For central data processing location, there are two possibilities: wired or wireless. Typically, wired systems employ glass fiber technology. Do IoT Fieldlab’s 5G expert Anup Bhattacharjee conducted a study on the scenarios for the communication infrastructure. He explains: “We compared all kinds of setups, varying with cloud and edge data storage solutions. Furthermore, we explored possible solutions using technologies such as 3G, 4G or 5G, and WiFi-P. For this project, we focused specifically on 5G technology and ‘beyond’ and the initial questions that arise when considering the use of 5G telecom networks for intelligent traffic control systems.”
Feedback from stakeholders
“After looking at the possible scenarios as to how 5G could potentially play a role in the intelligent traffic system applications, we discussed the same with the stakeholders of this domain such as government bodies, traffic consultants, traffic equipment manufacturers, connected vehicle manufacturer and telecom experts”, he adds. “We also received feedback during a meeting with to the Contactgroep Verkeersregel- technici Nederland (CVN) members. They emphasized that the reliability of communication links is crucial. Most municipalities currently rely on glass fiber for these installations, but when glass fiber is unavailable, 5G could be a viable alternative. A good understanding of the benefits and valuable advancements is paramount for the adoption of this new technology. Naturally, there is a lot of concern regarding privacy regulations (GDPR), which should be taken into account. In addition, system reliability is very important in the field of connected vehicles. The slightest disruption can cause critical sensor data to not be processed properly, posing risks. Even a minor hickup can translate to critical sensor data being missed during transmission, potentially causing risk. Therefore, in future studies, all these aspects have to be thoroughly investigated, in order to define the right conditions to involve information from connected vehicles in traffic light control.”
Sascha agrees: “Governments are interested in future technology, but we need the cooperation from all the different standpoints. And they obviously place high demands on the security and reliability of data exchange. The Netherlands is far ahead of the rest of the world in terms of knowledge and innovation. Scandinavia and Germany are also considerably developing traffic control tools. The TU Delft scientists and fieldlabs are keeping a close eye on all developments. We want to stay in the lead and will certainly continue to refine and improve the algorithms. With our continuous effort, the systems will soon be 5G-ready and we will be able to involve suppliers and governments in this development. I think it would be great to witness a stream of cyclists getting the green light at exactly the right time, and long enough, in a few years.”