How real-time monitoring could extend life of aging highway bridges

With limited maintenance budgets and a growing number of structurally aging bridges, transportation agencies are under pressure to prioritize repairs more precisely. A new study published in Infrastructures finds that real-time bridge health monitoring can provide the data needed to shift from reactive maintenance to risk-informed planning.

Titled Bridge Health Monitoring and Assessment in Industry 5.0: Lessons Learned from Long-Term Real-Time Field Monitoring of Highway Bridges, the study analyzes long-term monitoring systems deployed on active highway bridges.

Why traditional bridge inspections are no longer enough

Many bridges were designed decades ago for traffic volumes and loading conditions that no longer apply. At the same time, inspection schedules mandated by regulation typically require evaluations every two years, relying primarily on visual checks. These methods, the authors note, are time-consuming, expensive, and inherently limited in their ability to detect hidden damage or time-dependent behavior such as crack propagation, joint malfunction, or bearing degradation.

Recent regulatory updates in the United States have moved toward risk-based inspection frameworks and data-driven assessment, signaling a shift away from purely visual approaches. However, the study finds that adoption of real-time monitoring in practice remains limited. Much of the existing research has focused on algorithms and modeling techniques without addressing how reliable, continuous data can be collected and sustained on active highway bridges under real environmental conditions.

To bridge that gap, the research team developed and deployed a low-cost, real-time monitoring system focused on bridge expansion joints and bearings, components identified by transportation agencies as particularly vulnerable and costly to maintain. By monitoring three in-service bridges over extended periods, the study provides empirical insight into what works, what fails, and what must change for real-time monitoring to scale.

The authors emphasize that real-time monitoring is not intended to eliminate inspections but to complement them. Continuous data streams allow agencies to detect abnormal behavior early, prioritize interventions, and shift from reactive maintenance to predictive planning. This approach supports safer operations, reduces unnecessary closures, and extends the service life of critical infrastructure.

Lessons from long-term field deployment

The study analyses in detail the lessons learned from real-world deployment rather than controlled laboratory environments. The authors present a unified monitoring framework that spans planning, data collection, analytics, and decision-making, highlighting the need for coordination across engineering, computing, and operations.

Sensor selection emerged as a critical factor. While high-precision wired sensors offer excellent accuracy, their cost, power demands, and maintenance requirements make them impractical for large-scale, long-term deployment. The study found that ultrasonic distance sensors, when combined with appropriate signal processing, provided sufficiently accurate measurements for monitoring longitudinal bridge movements while remaining robust, low-cost, and energy efficient. Importantly, the authors stress that sensor selection must be based on field validation rather than manufacturer specifications alone.

Environmental durability proved equally important. Bridges expose equipment to moisture, deicing chemicals, temperature extremes, and mechanical vibration. Without proper enclosure design and corrosion resistance, sensors degrade quickly, compromising data quality. The research highlights the need for environmental hardening and adaptable mounting systems to ensure long-term stability.

Wireless communication was another decisive element. The study evaluated multiple networking options and identified low-power, scalable protocols as best suited for bridge monitoring. In particular, time-slotted wireless networks enabled reliable data transmission across multiple sensing nodes while maintaining energy efficiency. The authors note that communication strategies must be tailored to network size, data rate, and power availability rather than adopting one-size-fits-all solutions.

Power management remains one of the most significant barriers to uninterrupted monitoring. Many bridges lack access to grid power, and while solar energy offers a viable alternative, its reliability is highly dependent on weather, site geometry, and seasonal variation. The study documents frequent power interruptions caused by shading, snow accumulation, and battery limitations. These challenges underscore the need for multi-modal energy strategies, low-power system design, and remote power monitoring to maximize uptime.

Digital twins, AI, and human-centric decision-making

The study explores how analytics can transform raw sensor measurements into actionable insights. A key feature of the proposed framework is the integration of digital twins, statistical models, and artificial intelligence within a cloud-based analytics environment.

Digital twins in this context combine physics-based finite element models with real-time sensor data to provide a virtual representation of bridge behavior. These models establish baseline responses and help interpret deviations observed in the field. The authors demonstrate how calibrated simulations can also generate training data for AI models, addressing the challenge of limited labeled data from in-service structures.

AI was used to estimate bearing stiffness and assess component health, while simpler statistical models captured temperature-driven movement trends in expansion joints. The study finds that both approaches have value when applied judiciously. AI models offer predictive power but depend heavily on the quality and representativeness of training data. Statistical models provide transparency and interpretability but may struggle with nonlinear behavior. The authors argue that combining these methods yields more robust assessments than relying on any single technique.

Human–machine collaboration is a recurring theme. Rather than automating decisions, the system is designed to support engineers and inspectors through interactive dashboards and real-time visualization. Cloud-based platforms allow users to access data remotely, compare measured behavior with expected performance, and identify anomalies that warrant further investigation. This approach aligns with Industry 5.0's emphasis on augmenting human judgment rather than replacing it.

The study also examines the balance between cloud and edge computing. While edge processing can reduce data transmission, the authors found that performing analytics on low-power sensor nodes often increased energy consumption and reduced system reliability. In contrast, cloud-based processing offered greater flexibility, scalability, and power availability. The findings suggest that edge computing should be applied selectively, depending on sampling rates and power constraints.