cs.LGMar 30, 2026

Detecting the Unexpected: AI-Driven Anomaly Detection in Smart Bridge Monitoring

Authors: Rahul Jaiswal, Joakim Hellum, Halvor Heiberg

Organizations: Smart Sensor Systems AS, Oslo, Norway

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Bridges are critical components of national infrastructure and smart cities. Therefore, smart bridge monitoring is essential for ensuring public safety and preventing catastrophic failures or accidents. Traditional bridge monitoring methods rely heavily on human visual inspections, which are time-consuming and prone to subjectivity and error. This paper proposes an artificial intelligence (AI)-driven anomaly detection approach for smart bridge monitoring. Specifically, a simple machine learning (ML) model is developed using real-time sensor data collected by the iBridge sensor devices installed on a bridge in Norway. The proposed model is evaluated against different ML models. Experimental results demonstrate that the density-based spatial clustering of applications with noise (DBSCAN)-based model outperforms in accurately detecting the anomalous events (bridge accident). These findings indicate that the proposed model is well-suited for smart bridge monitoring and can enhance public safety by enabling the timely detection of unforeseen incidents.

Figures & tables

Fig. 1: A toy example of anomaly detection in sensor data.
Fig. 2: A simple illustration of an iTree structure.
Fig. 3: A simple illustration of an autoencoder.
Fig. 4: A simple illustration of the DBSCAN method.
Fig. 5: The proposed anomaly detection model.
Fig. 6: A simple illustration of the iBridge sensor device.
Fig. 7: Illustration of installed iBridge devices on the bridge.
Fig. 8: All features in the bridge monitoring dataset.
Fig. 9: acx_A and acx_B of the bridge monitoring dataset.
Fig. 10: All features except acx_A and acx_B of the dataset.
Fig. 11: Actual (ground-truth) accident timestamp.
ModelsParameters
Isolation Forestn_estimators 1 1 1 In Isolation Forest, n_estimators is the number of isolation trees, contamination is the expected anomaly proportion and n_jobs is the available CPU cores ( n_jobs =-1= all available). =1000, contamination=0.0001, n_jobs=-1
Autoencoderinput layer dimension=8,
 encoder input dimension=4, activation function=tanh,
 bottleneck layer dimension=1, activation function=linear,
 decoder input dimension=4, activation function=tanh,
 output layer dimension=8, activation function=linear,
TABLE I: Parameters used in each ML model.
Fig. 12: Detected anomalies using the Isolation Forest model.
Fig. 13: Detected anomalies using the autoencoder model.
Fig. 14: Detected anomalies using the DBSCAN model.
More in this topic

Anomaly Detection

Anomaly detection finds rare or unexpected patterns in data, from defective products and system faults to fraud and security threats.

See all 652 papers

Explore similar work

CardsList

  1. Smart Sensing for Safer Bridges: From Sensor Signals to AI-Driven Anomaly Detection

    Oct 2, 2026Rahul Jaiswal, Joakim Hellum, Halvor HeibergOnline Anomaly DetectionTime-Series Anomaly Detection

    Smart Sensor Systems AS Oslo, Norway

  2. A Vehicle-Integrated Approach to Digital Twin Deployment for Bridges Through Drive-By Sensing

    Sep 25, 2026Zihao Liu, Daigo Kawabe, Jiaji Wang +2Structural Health MonitoringDigital Twins

    School of Civil and Environmental Engineering, University of New South Wales, Sydney, Australia · KyoCenseo Inc., Kyoto, Japan · Department of Civil Engineering, The University of Hong Kong, Hong Kong, China +1

  3. Traffic and weather driven hybrid digital twin for bridge monitoring

    Mar 14, 2026Phani Raja Bharath Balijepalli, Bulent Soykan, Veeraraghava Raju HastiStructural Health MonitoringDigital Twins

    University of Central Florida, Orlando, FL, USA