Bridge structures play a critical role in transportation networks, yet continuous monitoring remains challenging due to the high cost and complexity of conventional structural health monitoring (SHM) systems. Accurate characterization of displacement distribution is essential for evaluating structural performance, detecting early deterioration, and enabling data-driven maintenance. To overcome cost and scalability barriers, we present an integrated sensing framework that combines distributed fiber-optic measurements from existing telecommunication cables with a low-cost radar–accelerometer fusion sensor capable of high-precision local displacement monitoring. The proposed compact device integrates a 60 GHz FMCW radar with a three-axis MEMS accelerometer to provide continuous displacement and acceleration measurements with sub-0.1 mm RMS accuracy. Constructed from commercially available components—including an Infineon millimeter-wave radar chipset and a microcontroller supporting Wi-Fi, Bluetooth, and Gigabit Ethernet—the system can be manufactured for under USD 1,000, enabling broad deployment across diverse bridge environments. A radar-target identification algorithm, informed by both radar and inertial data, mitigates multi-reflection artifacts and phase discontinuities inherent in FMCW measurements, ensuring reliable local sensing under field conditions. To expand the sensing coverage beyond a single point, pre-existing fiber-optic cables—originally installed for telecommunications—are repurposed as distributed strain sensors. Using distributed fiber-optic sensing, continuous strain and deformation profiles can be captured along long bridge spans without additional cabling or intrusive installation. All sensing modules, including the radar–accelerometer units and fiber-optic interrogators, are synchronized through a common NTP reference, enabling seamless fusion of localized displacement, global strain distribution, and dynamic response data into a unified SHM framework. By integrating high-precision local sensing with wide-area distributed fiber-optic measurements, the proposed system offers a cost-effective and scalable approach for capturing displacement distribution across bridge structures. This multi-modal framework provides a practical pathway for early anomaly detection, enhanced serviceability assessment, and long-term lifecycle management of aging or newly constructed bridges.

Highway pavement networks require frequent condition assessment, but conventional inspections with dedicated survey vehicles remain costly and are conducted at long intervals. This delays timely maintenance and increases life-cycle costs. This study presents an AI-based pavement condition assessment framework based on multi-modal sensor fusion, integrating video and vibration data from a vehicle-mounted smartphone with distributed fiber-optic sensing (FOS) installed along the roadside. These two data collection approaches are complementary, as the smartphone sensors cover the full network at low cost but produce noisier data, whereas FOS offers high-fidelity measurements only at equipped sections. Data from all three modalities were collected over multiple runs on a Korean expressway and evaluated across seven scenarios covering both individual and fused sensor configurations. Three tree-based classifiers (LightGBM, XGBoost, Random Forest) were trained to predict the International Roughness Index (IRI), a roughness-based index derived from the road's longitudinal profile, and the Highway Pavement Condition Index (HPCI), a composite index that incorporates both roughness and surface distress. The results indicate that vibration features contributed effectively to both IRI and HPCI by reflecting the vehicle's response to surface irregularities. Vision features contributed primarily to HPCI by capturing surface defects, while FOS features improved both predictions by providing structural response measurements inaccessible to the vehicle-mounted sensors. The findings demonstrate that combining modalities generally outperformed individual modalities, but the most effective combination differed by target index, offering practical guidance on how to pair sensing modalities for different pavement condition metrics.

Monitoring the condition and performance of bridges is critical for enhancing the resilience and safety of urban infrastructure. In particular, displacement is a critical measurement for many bridge health monitoring tasks such as load rating and serviceability assessment. Thus, many sensing modalities and sensing fusion methods have been developed for displacement estimation; however, such methods often require fixed reference points, are sensitive to occlusion, or costly to install and maintain, and thus not scalable at a city level.
Our previous works have demonstrated that we can convert pre-existing telecommunication fiber optic cables into a dense array of vibration sensors using distributed acoustic sensing (DAS). This enables scalable, high spatial resolution (meter-level) bridge vibration sensing, as a single interrogator can turn more than 100 km of telecommunication fiber into distributed strain sensors. Thus far, our group has used this to measure the dynamic component (1-10 Hz for short to medium-span bridges) of bridge vibrations and identify modal parameters. However, in order to accurately estimate displacement, we need to extract the scaled, quasi-static (0.1 – 1 Hz) component of the bridge vibration as well.
To this end, we develop a novel displacement estimation framework for distributed strain measured from non-dedicated telecommunication fiber optic cables. The key challenge is the coupling between the fiber and the bridge that distorts the bridge response measurements. The pre-existing telecommunication fiber optic cables are often laid inside conduits, which are attached to bridges. As the bridge vibration passes through these attachments, they introduce noise as well as frequency-dependent (for different bridge modes) amplitude distortion. Using double integration amplifies such noise, especially in the lower quasi-static band, resulting in low accuracy.
To address this, we introduce a structure-aware FIR filter to reconstruct bridge displacement from telecommunication fiber. We first address the higher noise in non-dedicated fiber with a finite impulse response (FIR) filter based on the regularized distributed strain-displacement relation. This suppresses the low spatial frequency drift and enables reliable estimation of the quasi-static components of bridge vibration. A transfer function parameterized for each mode band is then calculated during the calibration step to address the frequency-dependent distortion from the coupling structure. This framework requires no prior knowledge or assumptions about modal shapes, and only a temporary sensor needs to be deployed at calibration time to identify the transfer function.
Based on this framework, we estimate the transfer function between the distributed strain measurements from the telecommunication fiber and actual bridge response, using a temporarily placed reference accelerometer sensor for calibration. Our evaluation results from two real-world in-service highway bridges (Bakdal 2 and Miho River bridges in Korea) demonstrate that DAS can serve as a displacement sensor with sub-millimeter accuracy.

The increasing prevalence of high-speed rail places significant mechanical demands
on tunnel infrastructure. As trains enter and exit tunnels at high velocities, they gen-
erate shock waves that induce cyclic stresses on the concrete lining. These
stresses can persist beyond the train passage, potentially leading to fatigue-induced fail-
ures or the propagation of structural cracks over time. Therefore, continuous monitoring
of the strains induced by these shock waves is critical to ensure the safety and
durability of tunnel systems. Conventional pressure monitoring techniques, such as pres-
sure sensors and acoustic monitoring systems provide localized data but are limited in
spatial resolution, costly to install and maintain, and difficult to scale for long tunnels.
Furthermore, these systems often require direct installation at specific points of inter-
est, restricting their ability to provide a comprehensive understanding of shock
wave spatial distribution throughout the tunnel.
To this end, we propose a novel approach that leverages pre-existing telecommunication-
tion optical fiber cables, originally installed for station communication, to detect strains
induced by shock waves. Through the principle of Distributed Acoustic Sens-
ing (DAS), pre-existing telecommunication fiber cables can be transformed into dense
arrays of virtual strain sensors capable of detecting perturbations along their entire length
with meter-level resolution. However, the recorded strain signals include both structural
vibrations from train movement, shock waves generated in the air, and other ambient sources of vibration of tunnels.
To tackle this challenge, we isolate vibrations induced by the shock waves
through a frequency-velocity (f-v) filter. Since shock waves propagate through
air at the speed of sound, approximately 343 m/s under standard atmospheric conditions,
this known propagation velocity serves as a discriminative feature to isolate them from
structural vibrations induced by the train or tunnel infrastructure. By applying the f–v
filter, we effectively suppress non-acoustic components in the DAS signal and extract
shock waveforms with high spatial (meter-level resolution) and temporal resolution-
lution. We validated our approach in a real-world experiment in the Korea Rail tunnels of
length of approximately 735 m in South Korea. The proposed method allows us to turn
telecommunication optical fiber cables into cost-effective, high-resolution tunnel moni-
toring sensors by isolating strains induced from shock waves, offering potential indicators of structural fatigue loads in the tunnels.

This paper presents a pioneering application of Distributed Fibre Optic Sensing (DFOS) technology for the dynamic response monitoring of the Zeeland Bridge, a 5 km long cantilever-balanced prestressed concrete bridge. The study explores the feasibility and potential of using pre-installed single-mode telecommunication fibres, embedded within ducts inside the bridge structure, as a sensing network for large-scale structural health monitoring.

Spatially high-resolution dynamic strain measurements were performed under two distinct loading scenarios: (i) a controlled bridge closure during which a test vehicle was driven over the deck, and (ii) normal operational conditions under ambient traffic. These measurement campaigns allowed for the evaluation of the bridge’s dynamic behaviour under both controlled and real-world loading conditions.

The DFOS system successfully captured vibration signatures induced by both test and ambient loads, demonstrating its capability to detect dynamic responses over the entire monitored length of the structure. The results highlight the effectiveness of leveraging existing telecom fibre infrastructure for distributed, continuous, and minimally invasive monitoring, eliminating the need for additional sensor installation.

Ongoing data analysis focuses on identifying clear correlations between locally varying structural characteristics of the bridge and its dynamic response. These correlations are expected to provide valuable insights into the bridge’s dynamic properties and spatial variability, contributing to an improved understanding of its structural behaviour.

Overall, this study demonstrates the strong potential of DFOS technology combined with existing telecommunication fibres for the long-term monitoring of large-scale civil infrastructure. The proposed approach offers a promising pathway towards more efficient asset management, condition assessment, and maintenance strategies for ageing bridge structures.

Distributed optical fiber sensors (DOFS) have emerged as a powerful tool for the structural health monitoring (SHM) of civil infrastructure. This paper reports on an 18-year longitudinal study of a pedestrian bridge constructed from ultra high strength fiber reinforced concrete (UFC). DOFS were installed on the main girder’s lower surface in August 2007, prior to service. Although UFC is designed to be crack-free under service loads, detecting potential microscopic cracks is critical for validating design assumptions. However, such cracks are too small for visual inspection. To overcome these limitations, we developed a statistical evaluation framework that leverages the high-solution spatial data from DOFS, utilizing pattern recognition approach based on normalized cross-correlation (NCC). This method enables the automated identification of microscopic strain anomalies across the entire monitored length. Our results confirm that no cracks exceeding 20 µm have occurred over 18 years of service.