Candidate ID: DC6
Topic: Training AI health monitoring algorithms using large-scale synthetic data
Beneficiary: INFRAPLAN
My name is Lou. I recently completed a Master’s degree in Geography with a concentration in Remote Sensing at the University of Zurich (UZH) and am now a PhD researcher on the FOURIER project at Infra Plan Consulting d.o.o. I have a multidisciplinary background in remote sensing, GIS, photogrammetry, and machine/deep learning, developed through research roles at UZH and ETH Zurich and complemented by industry experience. In my master’s thesis, I quantified land surface temperature changes linked to land cover transitions, showing how deforestation and urbanization correspond to distinct global warming and cooling patterns.
My PhD research focuses on training AI-based infrastructure health monitoring algorithms using large-scale synthetic datasets to overcome the scarcity of labeled field data on damage and degradation events. The project generates synthetic multi-time series data by combining digital twin simulations grounded in engineering models with uncertainties, pre-stress effects, external loading, failure responses, and recovery. Generative adversarial networks (GANs) are then applied to scale these simulations, creating large datasets for training and validating AI algorithms that infer physical degradation parameters from sensor data.
I was drawn to the FOURIER project because it integrates AI, modeling, and monitoring to deliver practical, decision-relevant tools for resilient infrastructure management. Its interdisciplinary and application-oriented approach aligns with my research interests, geospatial analytics background, and focus on robust, data-driven methods for monitoring infrastructure under uncertainty.
