Establishing a Data‐Driven Pseudo‐Baseline for Bridge Monitoring Using ANN and Matrix Profiling

Abstract Structural Health Monitoring (SHM) is crucial for ensuring bridge safety, yet many methods rely on baseline data or known damage states—often unavailable for aging structures. To address this, we propose a new approach that combines Artificial Neural Networks (ANNs) with matrix profiling (MP) to create a “pseudo‐baseline” for predicting bridge behavior. Physics‐Informed Neural Networks (PINNs) incorporate physical laws into the model, while MP detects patterns and subtle anomalies in structural data. This method links structural responses, like strain and displacement, to environmental factors such as temperature and humidity. By analyzing these relationships, we can model normal bridge behavior without needing complete historical data. The approach is validated using performance metrics such as R 2 , Root Mean Square Error (RMSE), and residual analysis. Our combined method offers an innovative solution for real‐time anomaly detection, providing a more accurate and proactive tool for long‐term bridge monitoring.