The EU-funded ACUMEN research project aims to improve transport network efficiency and promote shared mobility options. The ACUMEN project partners are developing new analysis and management strategies to achieve this, with the Aimsun team chiefly involved in anomaly detection, network stress testing and digital integration.

Unpredicted and unexpected changes in mobility patterns can happen for many reasons, such as collisions, vehicle breakdowns, infrastructure failure, or extreme weather.

While there are ways to improve the resilience of critical infrastructure, it is just as important to react quickly to unexpected events to minimize their impact.  

Incident detection and mitigation

Aimsun, Aalto University, and the National Technical University of Athens have been working on algorithms that can detect unexpected events at unprecedented speeds and levels of precision thanks to recent advances in machine learning techniques, data and computation. These advances mean that unavoidable disruptions can be managed with minimal friction and delay to the road users.

The Aimsun team also brings commercial experience of how these algorithms need to perform to be market ready for real-life use cases. In addition, we can benchmark new methods against state-of-the-art Aimsun tools to measure their improvement. We are currently in the process of designing the methods, characteristics and specifications so we can understand their data needs and ensure they fit the project goals.

Infrastructure resilience

Following anomaly detection is the work on infrastructure resilience, or as we call it in ACUMEN, “network stress testing”. In a nutshell, this process checks how well the network keeps working when something fails.

This is, of course, an oversimplification as there are many variables that affect the outcome, such as the type of failure, its location, which modes were affected and how severely and for how long, etc. The key is to assess how much disruption different combinations of these different variables will cause.

In essence, this requires sensitivity analysis where the perturbations to initial conditions are different degrees of disruption at different locations. Sensitivity analysis can be summarized as a technique that enables detection of how much a change in input impacts the output; in the case of ACUMEN, we launch multiple simulations with different degrees of disruption – read this post by Alkis Papadoulis for more information on how Aimsun does sensitivity analysis.

This analysis helps us see which parts of the infrastructure have low resilience and need to be monitored closely, and which ones are more strongly backed up by other parts of the network, so that even under disruption, the network is able to operate at a good level of service. The University of Naples (UNINA) is undertak ing this part of the ACUMEN project, defining the experiments to be carried on the network, and using the sensitivity analysis pipeline generated by Aimsun.

Digital twin

Finally, our other main piece of ongoing work is to incorporate simulation outputs into a web-based, project-wide digital twin. This will show the main results of the project’s different use cases in a single, web-based interface. Aimsun will facilitate the integration of the different simulation outputs, such as vehicle trajectories, into the Digital Twin to be easily visualized by the different stakeholders involved in the project.

These strategies will be tested in ACUMEN’s four test cities of Amsterdam, Athens, Helsinki and Luxembourg, each of which focus on a different aspect of the project. In Helsinki for example, Aimsun is testing trajectory pricing and how to nudge people into choosing the right mode of transport to optimize the system. 

Watch this space for more updates and results, coming soon. https://acumen-project.eu/