Listening to bridges through everyday traffic: mobile sensing and AI for safer infrastructures
We are pleased to welcome our former colleague Kévin Theunissen, who will give a presentation on the research he has conducted in the United States this Thursday, September 10, 2026, from 10:30 a.m. to 12:00 p.m. in Room 0/429, Building B52.
His presentation is titled “Listening to bridges through everyday traffic: mobile sensing and AI for safer infrastructure.”
If you have any questions, please contact Margaux Geuzaine.
Abstract
A bridge cannot speak, but every vehicle crossing makes it vibrate. Like a doctor listening to a heartbeat, engineers can ask: How does the bridge move? Where does it feel stress? Which parts may become more vulnerable over time? These questions are central to bridge monitoring, especially as infrastructure ages, traBic increases, and maintenance decisions must be made with limited resources.
Mobile sensing oBers a new way to listen to bridges. Instead of relying only on sensors permanently installed on a structure, it uses measurements collected by sensors moving across it, such as accelerometers inside vehicles or smartphones. Each crossing becomes a short observation of the bridge’s dynamic behavior. When many passages are combined, everyday traBic can provide repeated, spatially rich information at a scale that fixed sensor networks often cannot reach. This makes mobile sensing especially attractive for large bridge inventories, long-term monitoring, and situations where installing dense instrumentation is too expensive or impractical.
However, mobile sensing also changes the engineering problem. The measured signal does not come from the bridge alone: it also contains vehicle dynamics, road roughness, speed eBects, sensor variability, and environmental noise. Each crossing captures only a partial and moving view of the structure, so the challenge is not only to measure vibrations, but to transform many imperfect observations into reliable structural information. This includes identifying vibration frequencies, understanding mode shapes, and eventually estimating quantities directly related to safety and durability.
Artificial intelligence provides a powerful layer for this transformation. AI can help assess signal quality, select useful passages, reduce noise, separate vehicle and bridge contributions, and learn mappings that are diBicult to describe with purely analytical models. More importantly, AI opens the door to virtual sensing: using data measured at one place, or from a moving sensor, to estimate structural responses at locations where no physical sensor is installed.
In this vision, mobile acceleration measurements can be transformed into fixed-reference bridge responses, including strain histories at critical locations. Strain is particularly important because it connects vibration monitoring to fatigue, damage accumulation, and life-cycle management. By combining mobile sensing with AI-based signal processing, everyday traBic can become more than a source of loading: it can
become a distributed diagnostic tool for understanding how bridges behave, how they age, and how we can maintain them more intelligently.
