Detecting the Unexpected: AI-Driven Anomaly Detection in Smart Bridge Monitoring ∗ Corresponding author PubID: pubid: 979-8-3315-3672-5/26/$31.00 ©2026 IEEE

Rahul Jaiswal ∗ , Joakim Hellum and Halvor Heiberg · 2026

Read Detecting the Unexpected: AI-Driven Anomaly Detection in Smart Bridge Monitoring ∗ Corresponding author PubID: pubid: 979-8-3315-3672-5/26/$31.00 ©2026 IEEE by Rahul Jaiswal ∗ , Joakim Hellum and Halvor Heiberg (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-28225.txt.

Introduction

Jaiswal, Hellum, and Heiberg use sensor data and machine learning to spot accidents on a smart bridge in Norway. Human inspections are slow and can miss or misjudge damage. Their system learns from live iBridge sensor streams and flags odd events as they happen. A density-based clustering model, DBSCAN, beat the rival models they tried at catching the accident case. The setup could warn operators fast when something goes wrong.

Catalog record

CREATORRahul Jaiswal ∗ , Joakim Hellum and Halvor Heiberg
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE23559
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.28225

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Detecting the Unexpected: AI-Driven Anomaly Detection in Smart Bridge Monitoring ∗ Corresponding author PubID: pubid: 979-8-3315-3672-5/26/$31.00 ©2026 IEEE Rahul Jaiswal ∗ , Joakim Hellum and Halvor Heiberg 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. I Introduction Bridges are the critical component of smart cities, forming the backbone of modern transportation systems. With hundreds of thousands of bridges deployed worldwide, for instance, over 18,000 in Norway [ 16 ] , ensuring their safety and reliability is of paramount importance. Traditionally, the assessment of bridge conditions has relied heavily on human visual inspections. However, such inspections are often time-consuming and prone to error, as accurately evaluating the extent of structural deterioration, particularly in concrete/wooden bridges, is challenging due to the limitations of human visual observations. To overcome these challenges, intelligent sensor-based, data-driven machine-learning bridge monitoring techniques have emerged. It enables timely assessment of structural integrity, enhancing pedestrian and public safety, and mitigating the risk of catastrophic failures or accidents. Effective bridge monitoring requires designing a smart model that can capture variations between normal and anomalous sensor data. Anomaly detection is an important application of data mining [ 9 ] and is commonly referred to as outlier detection. In the context of bridge sensor data, anomalies or outliers correspond to the measurements that deviate from the normal signal or data patterns. These deviations may arise from structural damage, abnormal loading conditions, environmental effects, or sensor malfunctions. The objective of anomaly detection is therefore to identify such inconsistent or irregular sensor observations that differ significantly from typical bridge behaviour. A toy example illustrating anomaly detection in sensor data is presented in Fig. 1 . As shown, the anomalous data point exhibits a significantly higher peak compared to the normal sensor measurements. Fig. 1: A toy example of anomaly detection in sensor data. Several approaches have been proposed in the literature for detecting anomalies in sensor data. Statistical-based methods [ 15 ] , such as multivariate models [ 5 ] and mean-variance-based techniques [ 7 ] , typically learn discriminative features from historical monitoring data representing normal structural behaviour. A data point is then classified as anomalous when it deviates significantly from these learned features. However, the effectiveness of statistical methods relies on the prior empirical knowledge, including assumptions about data distributions and model parameters used for anomaly identification. In practice, the true distribution of real-world sensor data is often unknown, which can limit the accuracy of anomaly detection using statistical methods. Recently, artificial intelligence (AI)-driven, particularly machine learning (ML)-based, approaches have gained significant attention for anomaly detection [ 14 ] , as they automatically extract discriminative features from raw bridge sensor data to train models. Neural network-based methods are also effective due to their strong representation and universal approximation capabilities [ 12 , 13 ] . For instance, a K-nearest neighbour-based anomaly detection framework is proposed in [ 8 ] using raw vibration measurements collected from a bridge. A convolutional neural network (CNN) is used in [ 1 ] to detect structural damages in bridge systems. An autoencoder is employed in [ 17 ] for anomaly detection in bridge structural health monitoring. The review of existing literature highlights that anomaly detection in bridge systems is critical for maintaining the structural performance of the bridge, ensuring pedestrian and public safety, and mitigating the risk of catastrophic failures or accidents. Motivated by these findings, this work focuses on machine learning-based anomaly detection for smart bridge monitoring. Specifically, we develop an efficient ML-based anomaly detection framework to identify a bridge accident captured by iBridge sensor devices (see Section IV-A ) installed on a bridge in Norway. The proposed model can enable bridge monitoring authorities to promptly detect anomalous events and take timely corrective actions, thereby reducing potential risks and preventing further adverse consequences. The remainder of this paper is organized as follows. Section II describes the machine learning methods employed for anomaly detection. Section III introduces the proposed anomaly detection model. Section IV describes the experimental dataset. Section V discusses the results, and Section VI concludes the paper and outlines directions for future work. II Background This section describes various machine learning techniques used in our experimental study for detecting anomalies in the bridge monitoring system. II-A Isolation Forest The Isolation Forest (IF) algorithm detects anomalies using an ensemble of binary trees [ 3 ] . In bridge monitoring applications, sensor data are isolated by randomly selecting a feature and an associated split value at each node. The algorithm constructs a forest, referred to as an iForest, comprising multiple isolation trees (iTrees) [ 19 ] . During training, the sensor data are recursively partitioned until the iTrees effectively separate normal data points from anomalous data points. Typically, anomalous data points are isolated near the root nodes of the iTrees, whereas normal operating conditions (points) require deeper traversals and are located farther from the root nodes. A simplified structure of an iTree [ 14 ] is illustrated in Fig. 2 . Here, external nodes have no children, whereas internal nodes contain two child nodes. In this representation, anomalies are typically associated with external nodes. Fig. 2: A simple illustration of an iTree structure. II-B Autoencoder An autoencoder (AE) [ 2 ] is a neural network designed to learn a compact (latent) representation of input data while minimizing the error in reconstructing the original signal. In bridge monitoring applications, autoencoders are typically trained using sensor measurement data that capture normal signal behaviour. Consequently, when anomalous sensor measurements, such as those caused by structural damage, a vehicular accident, an unusual load on the bridge, or sensor malfunctions, are encountered, the reconstruction error increases significantly. This reconstruction error can therefore be used as an effective indicator for anomaly detection. Fig. 3: A simple illustration of an autoencoder. An autoencoder consists of an encoder and a decoder. The encoder compresses input data into a low-dimensional latent representation (also called a bottleneck), and the decoder reconstructs the original input from that latent representation, as shown in Fig. 3 . The autoencoder model is trained by minimizing the reconstruction error between the input and its reconstruction. Next, we briefly present the mathematical formulation of an autoencoder. Let a sensor device with d channels generate measurements over time. At time t , the sensor measurements are given as: \mathbf{x}_{t}=x_{t}^{(1)},x_{t}^{(2)},\dots,x_{t}^{(d)} (1) where, x_{t}^{(i)} denotes the measurements of the i-th sensor device. The encoder compresses the input into a low-dimensional latent

Read the complete text