New research can reduce traffic congestion

Researchers have discovered that traffic flow and chemical reaction models can be mathematically described in the same way. Mathematical models commonly used in chemistry to describe how molecules react to produce substances, may turn out to have implications for mimicking how the density of vehicles change in transport networks. Balázs Kulcsár, professor at the Department of Electrical Engineering, explains:

“Chemical reactions happen all the time. Carbonic acid is for example formed in soda drinks when carbon dioxide is pushed into water with soda machines. We have discovered that the same mathematical models can be used to describe chemical reactions as can be used to model traffic networks. The results have multiple benefits for each of the fields chemistry, traffic flow theory and applied mathematics. “

Mathematics as a springboard

Within material science, optimal and energy-efficient management of "molecular crowding" (concentrations) can help create optimal and finely structured materials using techniques e.g. borrowed from traffic light control. In traffic flow theory, such models can help us understand and predict the behavior of large-scale traffic networks and balance possible traffic congestion. Traffic flow theorists can learn these new tricks and concepts from chemistry. With the help of mathematics, the two seemingly disparate fields of chemistry and transportation science can be brought together with interpretable patterns.

Annika Lang, professor at the Division of Applied Mathematics and Statistics, elaborates:

“In chemistry, to the best of our knowledge, there has never been a published method for breaking traffic jams of molecules using traffic control paradigms. The concept of molecules behaving like flowing vehicles is a modern view. With the help of mathematics, and more specifically partial differential equations and the approximation of their solutions, we are now bringing these two seemingly distant sciences together and discovering new connections.”

Ready to implement 

The results could mean more accurate congestion models and more reliable traffic flow prediction tools. The scientists have gathered real world data for the model using pure data from highways. A model-based congestion control with higher accuracy could reduce the average time spent in congestion through, for example, speed limiting. Yet, for a final implementation, legal and policy aspects would also need to be solved - something which can vary on country legislation.

Balázs concludes:

“When we discovered that large collections of molecules and cars behave mathematically in the same way, it was somewhat of “an eureka moment”. Five years ago, a researcher in bioinformatics and chemistry and a researcher from transport science realized the similarities. The occurrence of 'chemical reactions' roughly means that vehicles can move forward on the road, since there is enough available space in front of them. Interestingly, the corresponding dynamics known from physical chemistry can be completely matched to the one used in traffic modeling. We are not only excited about the discovery and the potential but also, we are very proud of the great international and interdisciplinary team.”

The research has been taking place at Chalmers since 2021 with the support of Chalmers Transport Area of Advance, Chalmers Artificial Intelligence Research Center, CHAIR and own faculty funds. The collaboration is carried out by Chalmers researchers Professor Balázs Kulcsár at the Department of Electrical Engineering and Professor Annika Lang at the Department of Mathematical Sciences together with researchers Assistant Professor Mike Pereira (Mines Paris - PSL University, France) and Professor Gábor Szederkényi (Pázmány Péter Catholic University, Hungary).