New Publication in Building and Environment: Reconstructing Indoor Airflow from Sparse Measurements

We are pleased to share our latest research paper on airflow modeling:

“A physics-informed Gaussian Markov random field framework for indoor airflow field estimation“
Building and Environment, 2026.
DOI: 10.1016/j.buildenv.2026.113264

Understanding indoor airflow patterns is essential for assessing air quality, thermal comfort, and pollutant dispersion. However, obtaining high-resolution airflow maps in real time remains a challenge: high-fidelity simulations (like CFD) are computationally demanding, while physical sensor networks only provide sparse, localized measurements. Within our research line on artificial olfaction and mobile robotics, this work presents a physics-guided Gaussian Markov Random Field framework (GMRF-W). By integrating fundamental fluid transport principles directly into a probabilistic graph, the proposed method reconstructs continuous 2D wind velocity fields from sparse measurements with high spatial accuracy and low computational cost.