Our mobility ecosystem is becoming increasingly complex, especially in urban areas. This makes it virtually impossible to manage traffic in the traditional centralized way. Is another approach conceivable? In the EU project DIT4TraM we apply swarm intelligence: with little or no control from above, yet still achieving a common goal. This approach is tested in six different pilots.
It’s a fascinating spectacle—tens of thousands of starlings that form beautiful moving patterns in the air. Their aim is to scare off their natural enemies, biologists explain. The great thing is that they succeed in doing so without any direction or guidance from above. How do they do that? The swarm creates harmonious patterns because all birds individually follow a small set of local rules. These rules are also remarkably simple, like ‘keep enough distance from the neighbors, but not too much.’
Swarm intelligence
Traffic participants also have a common interest, namely a smooth and safe flow of traffic. To achieve these goals, road authorities use centralized traffic and mobility management to control and correct the traffic flows. But our mobility ecosystem has become so complex over the years that controlling traffic ‘from above’ is no longer tenable—especially if we want to manage it in a multi-modal way and include new transport services as well.
This forces us to think about a completely different approach to our traffic and mobility management. In this context, can we learn from the starlings and stop regulating centrally and top-down, and instead start regulating locally and bottom-up?
In the DIT4TraM project, Distributed Intelligence and Technology for Traffic and Mobility Management, twenty knowledge institutions, companies and governments are investigating and testing for the first time whether this ‘swarm intelligence’ approach is feasible for traffic and transport. Our focus is on individual travelers, connected cars, smart bicycles, and intelligent traffic control systems. Simply put, we are exploring how we can get these agents to communicate and interact locally in such a way that they automatically contribute to the greater goal of a smooth and safe traffic flow.
Four applications
DIT4TraM runs from 1 September 2021 to 1 September 2024. Our goal is to develop control concepts and algorithms based on swarm intelligence for the widest possible range of applications during this project period. We’re including different modalities, from pedestrians to cars to subsystems. And we focus on different use cases, from local traffic control to regional traffic and mobility management.
We are, of course, looking at fully decentralized solutions with one hundred percent self-organization, drawn up according to the so-called ‘mechanism design’ concept. But we also include solutions with distributed intelligence: solutions that are local where possible and centralized where necessary. In this case, central intelligence, such as a traffic management system at the traffic center, monitors the situation in general and only interferes if the local intelligence falls short.
Six pilots
To test and develop all new knowledge and insights in practice, six pilot studies in four EU countries are organized: in France (Bordeaux), Greece (Glyfada and Athens), the Netherlands (Amsterdam and Utrecht), and Spain (AP-7 freeway). With these pilots we test the gains to be achieved for traffic flow, safety and liveability and look at the effects on reliability and resilience.
With these results, we create a better understanding of which principles of swarm intelligence can be used for traffic and transport. And this knowledge is expected to lay the foundation for a 180-degree paradigm shift in our traffic and mobility management.