Reviewing a Decade of Structural Health Monitoring in Footbridges: Advances, Challenges, and Future Directions
1
Centre for Infrastructure Engineering, Western Sydney University, Locked Bag 1797, Penrith, NSW 2751, Australia
2
School of Engineering, Design & Built Environment, Western Sydney University, Locked Bag 1797, Penrith, NSW 2751, Australia
3
Zhejiang University-University of Illinois Urbana-Champaign Institute, Zhejiang University, Hangzhou 310027, China
4
School of Business, Torrens University Australia, 1/37 Foveaux St, Surry Hills, NSW 2010, Australia
*
Author to whom correspondence should be addressed.
Remote Sens. 2025, 17(16), 2807; https://doi.org/10.3390/rs17162807
Submission received: 13 June 2025
/
Revised: 31 July 2025
/
Accepted: 8 August 2025
/
Published: 13 August 2025
(This article belongs to the Special Issue Advances in Remote Sensing and Digital Twin Technologies for Transportation Infrastructure)
Abstract
Aging infrastructure is a growing concern worldwide, with many bridges exceeding 50 years of service, prompting questions about their structural integrity. Over the past decade, the deterioration of bridges has driven extensive research into Structural Health Monitoring (SHM), a tool for early detection of structural deterioration, with particular emphasis on remote-sensing technologies. This review combines a scientometric analysis and a state-of-the-art review to assess recent advancements in the field. From a dataset of 702 publications (2014–2024), 171 relevant papers were analyzed, covering key SHM aspects including sensing devices, data acquisition, processing, damage detection, and reporting. Results show a 433% increase in publications, with the United States leading in output (28.65%), and Glisic, B., with collaborators forming the largest research cluster (11.7%). Accelerometers are the most commonly used sensors (50.88%), and data processing dominates the research focus (50.29%). Key challenges identified include cost (noted in 17.5% of studies), data corruption, and WSN limitations, particularly energy supply. Trends show a notable growth in AI applications (400%), and increasing interest in low-cost, crowdsource-based SHM using smartphones, MEMS, and cameras. These findings highlight both progress and future opportunities in SHM of footbridges.
1. Introduction
The American Society of Civil Engineering reported that there are around 617 thousand (approximately 42%) bridges in America that are older than 50 years [1]. Additionally, 46,154 bridges are classified as structurally deficient [1]. In New South Wales, Australia, approximately 70% of bridges were constructed before 1985, with a significant proportion dating back to the mid-1930s and a construction peak occurring in the 1970s [2]. Furthermore, a study by Salvatore et al. [3], which analyzed a dataset of 447 bridges uniformly distributed over the Italian territory, found that 34% of them exhibited a high or medium-high level of structural-foundational risk. Replacing these structures would entail substantial costs, demand extensive manual labor, and exceed available financial and human resources [4]. As a result, researching SHM of bridges is vital for early detection of structural reliability, ensuring both safety and economic viability.
A SHM system is an invaluable tool to identify aging structures and damages, aiding manufacturing maintenance, transportation, and building assets that have the potential to extend the structural life, enhance the safety of the public, and significantly minimize restoration costs [5]. SHM relies on sensors to collect accurate and continuous real-time measurements of structural conditions, transmitting this information to a central system or server that can analyze, compare and provide warnings should an irregular pattern occur. This automated real-time monitoring of footbridges could potentially make improvements in the effectiveness, speed and expenditure involved in bridge inspections compared to manual inspection methods that may require specialist inspectors that are time-consuming, expensive, and subjective [6].
SHM for bridges, including pedestrian bridges, railway bridges, road bridges, and road-rail bridges, is a broad field of study. Pedestrian bridges and vehicular bridges (road traffic and rail) vary in several key aspects. Pedestrian bridges are often more sensitive to minor vibrations attributed to resonance from the steps produced by pedestrians to the natural frequencies of the bridge [7], are less complex in design, have shorter spans, and are more responsive to ambient excitations [8]. Pedestrian bridges are particularly well-suited for testing emerging SHM technologies, primarily due to their lower associated risks, ease of access, and more straightforward approval processes from the road authority compared to vehicular bridges. For these reasons, this review has chosen to focus specifically on footbridges.
The number of research papers collected from this study has suggested that in the last 10 years, the studies on footbridge SHM employing various sensing technologies and methodologies have grown, which implies that real-time automated inspection of bridges [9] and the importance of footbridge SHM has received attention from researchers and infrastructure owners, particularly in the area of sensor development, damage detection, and data processing and analysis.
This study seeks to present an overview of SHM of footbridges via a combination of scientometric analysis together with state-of-the-art review. The objectives and aims of this review are exploring and classifying various categories of SHM for footbridges, highlighting advancements, identifying research gaps and challenges, and suggesting future directions. This review collected over 702 research papers through literature searches from 2014 to 2024. Through quantitative analysis and comprehensive discussion of the accumulated research papers, this study identifies advancements, research gaps, and challenges, followed by the proposal of future directions.
The remaining sections have the following structure: Section 2 presents the scientometric review that covers the literature search methodology, co-occurrence keyword analysis, co-authorship analysis, regions analysis, sensors count and popularity and data transmission medium. Section 3 expands into a state-of-the-art review by examining each collected paper. Section 4 looks into the challenges faced by the researchers and what directions are likely to move in the field, and finally, Section 5 provides concluding remarks.
2. Scientometric Review
A comprehensive examination of literature in a specific field that primarily uses quantitative study and statistical analysis to examine the influence of journals, institutions and networks in a targeted field describes a scientometric review. This review method typically emphasizes the study of literature patterns, co-citation, co-writers, and research efficiency and effectiveness, which are crucial for uncovering directions, determining significant research, and discovering how knowledge evolves within a specific field [10].
This research proposes a quantitative study of published literature related to footbridges and SHM. This methodology employs bibliometric techniques on published literature to analyze the structure, dynamics, and impact using extensive academic datasets [10]. The visualization offers a representation of the development and transformation of the studied field throughout the reviewed period.
Key researchers, publications, and countries essential to this study are identified using VOSviewer software and custom-developed R scripts generating reports using co-authorship author analysis, keyword co-occurrence and co-authorship countries investigation.
2.1. Methodology Use in Literature Search
Within the scope of this study, from 2014 to 2024, there was a growth in literature circulated in journals and proceedings of the conferences. Scopus, which contains 90 million records published by 7000 publishers in 105 countries [11], serves as the key data source for this review. Web of Science, Engineering Village, and Institute of Electrical and Electronics Engineers (IEEE) Xplore were the secondary sources of information. The literature search was conducted in four phases: identification, screening, eligibility, and finally, inclusion. Figure 1 illustrates the procedure’s flow used during the literature search. The literature search keywords consist of Boolean Operators: (“structural health monitoring” OR SHM) AND (footbridge OR “pedestrian bridge”). The identification phase collected a total of 702 papers restricted within the years 2014 to 2024. The screening phase removed 440 papers due to duplication, with 262 papers left. After manually reviewing and verifying the abstracts of the 262 papers, we discovered that although they matched the search criteria, some were not actually relevant to footbridges or SHM. These papers merely referenced terms like “footbridge” or “pedestrian bridge” or “SHM” in their text, leading to their inclusion by the data source provider. As a result, a further 91 papers were excluded.
Finally, the review analyses the remaining 171 papers to determine their objectives. Figure 2 illustrates the historical pattern of research publications in the interest field. During the last ten years, the number of publications increased, reaching 32 in 2024, suggesting growth in the field of footbridge SHM. Table 1 presents the proportions of publication types, with journal articles covering 53.8%, conference papers 44.4%, and book sections 1.8%.
Table 2 and Table 3 present the topmost reputable journals and conferences generated using a custom-developed R script. Table 2 reveals that the Journal of Civil Structural Health Monitoring published the most papers related to footbridge SHM among journals. In Table 3, the European Workshop on SHM published the most papers among conferences.
2.2. Co-Occurrence of Keyword Analysis
Keywords are crucial terms representing an academic study’s research focus and theoretical background, allowing readers to understand the article’s content in a brief and providing valuable information for researchers searching for related articles or conducting surveys on specific topics [12]. The VOSviewer software generated a network map of co-occurring keywords after analyzing and evaluating the keywords of the published papers from the dataset. This network map displays the keywords in a spatially arranged, distance-based diagram. In a distance-based diagram, nodes in a bibliometric network are positioned so that the distance between them indicates their relatedness, with smaller distances representing higher relatedness [13]. Each note represents a keyword, and each link represents a link in this type of distance-based network diagram [14]. Figure 3 shows a total of 33 nodes and 656 links. Table 4 shows details of each node corresponding to the headers of the table.
Table 4 indicates that the primary keywords are “structural health monitoring” and “footbridges”, suggesting substantial research in this area. A detailed analysis of the table data reveals that keywords like “computer vision”, “wireless sensor networks”, and “neural networks” related to footbridge SHM are sparse, indicating a crucial need for further research in these areas.
2.3. Analysis of Co-Authorship
The co-authorship network created using bibliometric techniques and VOSviewer software illustrates the research contributions of key authors and the collaborations among various authors in the field. The network shows the number of literature presented by nodes and the relationship with other authors presented by links. The size of the nodes is relative to the publications count of that author. Figure 4 presents a diagram of the largest cluster within the connected network.
R script custom-developed in this research extracted the bibliography data to produce a comma-separated values file with the list of the most productive authors presented in Table 5. From this table, researchers from Princeton University lead the top 1st and 3rd positions starting from Glisic, B. and Abdel-Jaber, H. The figure and table presented in the co-authorship analysis would be helpful information for researchers to identify key authors and organizations in the field of SHM in footbridges.
2.4. Network of Countries/Regions
Utilizing the VOSviewer software, this research produced a network that illustrates the distribution of research publications based on the contributions of different countries, as illustrated in Figure 5. Based on the data in Table 6, the countries that contributed to the most publications in SHM in footbridges are the United States, China, Italy, the United Kingdom, and Poland.
2.5. Sensor Count, Popularity of Sensor, and Data Transmission Medium
This review examined each publication, extracted the number of sensors used, types of sensors and data communication medium and recorded them in a spreadsheet. As some publications did not specify the sensors used in their studies, the tables presented in this section include only the sensors explicitly described in the collected publications. Table 7 reveals the most used sensors, Table 8 lists the sensor count employed, and Table 9 shows the communication medium described in the publications.
3. State-of-the-Art Review
After completing the scientometric overview, this study continues the methodology by expanding into a state-of-the-art review by thoroughly examining each collected paper and grouping them into categories that best represent their primary focus.
The seven main categories identified from these papers are planning and workflow, sensors, data acquisition, data transmission, data processing and analysis, damage detection, and reporting. Table 10 shows the number of papers in each category. Data transmission was not found in the included publications as the core research activity; however, due to the importance of transmission of data to the SHM system, it is included as a main category here and will then be further discussed in Section 4.
The study further expanded the analysis of the major categories by examining the purpose of each paper, thereby identifying subcategories within the collected research papers. Figure 6 shows the results on the subcategories identified.
3.1. Planning and Workflow
Effective planning and comprehensive documentation of the SHM workflow lifecycle are vital. Researchers in this area have proposed methods like 2D imaging, 3D modeling, Finite Element Analysis (FEA), Virtual Reality (VR), and Operational Modal Analysis (OMA) to support documentation, planning, and virtual site visits for asset stakeholders. Based on the data collected, the planning and workflow category is the least researched area, covering only 1.8% of the included publications.
Virtual Reality
In two published studies, Napolitano et al. [16] and Napolitano et al. [17] proposed a method to enhance current workflow practices for long-term monitoring and assessment. The methodology presented by Napolitano et al. [16] outlines a digital workflow for organizing and incorporating structural data and documentation into a virtual reality environment to facilitate communication between disparate groups. Using 360-degree spherical imaging (Ricoh Theta S camera) and virtual environment software (Kolor Panotour Pro) to stitch together a virtual environment, Napolitano et al. [16] implemented this methodology on Streicker Bridge, a 250-foot pedestrian bridge, in Princeton, NJ, USA. The sensors, strain sensors, temperature sensors, displacement sensors, and sensing sheet specimens fitted on the bridge were embedded into a 3D model [17]. By integrating VR with information modeling (IM), Napolitano et al. [17] determined that this methodology can improve communication in 3D SHM projects.
Working in the same area of VR method development, Luleci et al. [18] proposed Light Detection and Ranging (LiDAR) combined with Unmanned Aerial Vehicle (UAV) photogrammetry for capturing the bridge to produce a VR model that generates 3D data points and meshed models. Aerial images captured were processed using Reality Capture® software to generate a 3D meshed model. Vibration data collected by Luleci et al. [18] from a 10-channel dynamic analyzer is processed further in MATLAB® with the use of Stochastic Subspace Identification (SSI) and covariance techniques to draw out the model parameters. With the use of SAP2000® FEA software, the resulting FEA was compared with OMA to study the anticipated configuration from the FEA and the actual behavior of the OMA. The outcome of Luleci et al. [18]’s research is a method in a VR environment that uses Unity software and Oculus Quest 2 headset to conduct “virtual visits” of bridge structure for assessment.
3.2. Sensors
Research papers on sensors contributed to the second-largest category from the total included publications, 19.9%, as seen in Table 10. Research papers related to sensors were then reviewed and subcategorized into energy harvesting (1 paper), method validation (4 papers),sensor deployment (1 paper), sensor development (25 papers), sensor fault detection (1 paper), and finally, sensor validation (2 papers).
3.2.1. Energy Harvesting
Integrating remote network monitoring with energy harvesting systems is an emerging field with promising applications for SHM in civil engineering infrastructures [19]. According to Jiménez-Alonso et al. [19], which conducted a feasibility study on harvesting energy from pedestrian vibration in footbridges, pedestrian-induced vibration energy harvesting is a new field. Two different harvesters, patch piezoelectric energy harvester (PPEH) and cantilevered piezoelectric energy harvester (CPEH), manufactured using different piezoelectric material Lead Zirconate Titanate (PZT) and Polyvinylidene Fluoride (PVDF) analyzed by Jiménez-Alonso et al. [19], produced good performance without compromising the structure’s comfort rating. Figure 7 shows the experiment’s results. Jiménez-Alonso et al. [19] documented that PVDF performed better than PZT regardless of the type of harvester, followed by the performance of CPEH, which is better than PPEH, irrespective of the material. Even though the analysis showed good performance, Jiménez-Alonso et al. [19] suggested that further studies are needed experimentally to assess this performance when they are used to generate electricity from pedestrian-induced vibrations of footbridges.
3.2.2. Method Validation
Method validation is a key requirement in performing any analytical process. In the method validation sub-category, this review examines publications focused on validating methodologies.
Distributed Fiber Optic Sensing (DFOS)
The use of DFOS for bridge monitoring remains a relatively new approach [20]. The DFOS sensing technique enables the measurement of strain, indirect derivation of displacement or deflection, and temperature changes throughout the entire span of a continuous structure [20]. In Sieńko et al. [20]’s research, the methodology of a SHM system of a pedestrian footbridge in Nowy Sącz installed with DFOS strain sensors called EpsilonRebar embedded within composite deck panels before concreting was validated. Results of the actual measurements from the EpsilonRebar when compared qualitatively with FEM generated with the use of SOFiSTiK software, proven to be exemplary [20].
Bednarski et al. [21] discussed the research on the temperature behavior of monolithic DFOS sensors established from theories and laboratory tests on different structures. The results indicated the linearity of the temperature behavior of the DFOS sensing device and validated that they are effective in specifying longitudinal strain and stress.
Smartphone Sensing Method Validation
With the use of both stationary and mobile smartphones, McSweeney et al. [22] aimed to validate the method of two mobility approaches for identifying bridge modal frequencies. The results indicated the potential of obtaining the bridge’s fundamental frequency for the identification process with fixed and moving smartphone measurements but faced operational indecisiveness with the mobile measurements.
In Komarizadehasl et al. [23]’s natural frequency analysis, the iPhone XR smartphone demonstrated the ability to gauge the bridge’s frequencies using the Phyphox app for data acquisition. In comparison to the reference device LARA, the iPhone XR demonstrated less than a 0.49% difference; however, it failed to detect two of the bridge’s longitudinal natural frequencies [23]. Despite these limitations, Komarizadehasl et al. [23] concluded that the built-in accelerometers in smartphones offer a significantly more convenient alternative to custom-developed systems.
Vehicle-Based Interaction Sensing
Aiming to overcome the traditional issues in OMA methods, Fiandaca et al. [24] proposed an innovative Vehicle-Based Interaction (VBI) method using accelerometers mounted on an electric mobile platform. When validated against a traditional OMA method, the VBI method showed a maximum discrepancy of 3%. Fiandaca et al. [24] concluded that the VBI method is reliable and cost-effective for identifying dynamic parameters.
3.2.3. Sensor Deployment
Sensor deployment is a critical challenge in sensor networks. Research on the sensor deployment problem focuses on positioning sensors efficiently to cover designated areas or specific points [25]. The reviewed research publications illustrated in Figure 6 indicate that sensor deployment for SHM of footbridges is a field with limited studies.
Unmanned Aerial Vehicle Sensors Deployment
Presently, SHM systems encounter challenges associated with costs and deployment difficulty [26]. Structures in remote or hard-to-access locations, such as suspension bridges and high-voltage pylons, are often situated in challenging terrain, making conventional access difficult and unsafe [27]. Therefore, the potential to deploy sensors on such infrastructure using autonomous vehicles such as un-crewed aerial vehicles (UAV) can be rewarding. In Satme et al. [27]’s research, UAV was used to deploy wireless sensor networks (WSN) on a footbridge at hard-to-reach locations due to their high mobility capability. A WSN sensor node with ARM Cortex-M7 microcontroller, high-performance SCA-3300 microelectromechanical systems (MEMS) accelerometer and an NRF24L01+ as the radio frequency communication module enclosed in an in-housed 3D printed PLA frame was developed. According to Satme et al. [27]’s research, the key to the remote deployment was the incorporation of the EPM-V3R5C electropermanent magnets controlled by the activation of magnetization and demagnetization during the flight on the UAV. Using this sensor deployment method, Satme et al. [27] reported the ability to capture extensive data and detect slight structural behaviors that are difficult to observe compared with the conventional sensor deployment methods.
3.2.4. Sensor Development
The sensor development is the largest sub-category in the sensors research category of the included publications. In this sub-category, this review examines research focused on sensor development.
Accelerometer Sensor
Vibration-based condition monitoring, which emphasizes system identification and modal analysis, has gained significant attention among widely used SHM techniques [28]. The development of accelerometer sensors emerged as the second-largest research subcategory within the sensor development field, according to the literature reviewed by this investigation. As shown in Table 11, 80% of the studies in this area utilized wireless networks for data transmission and low-cost MEMS accelerometers as the sensor.
Cruz et al. [29] proposed a WSN that includes MEMS accelerometers acquainted by WSN nodes for fast speed acquisition of movements of a footbridge. Cruz et al. [29] used a function generator to calibrate and evaluate the accelerometers’ precision. Cruz et al. [29] stored the recorded vibration data in a remote cloud-based platform and further processed the data with the fast Fourier transform function in MATLAB software [29]. Three WSN nodes installed at the girder of a pedestrian bridge in Manila formed the pilot test system, according to Cruz et al. [29]’s report. The experiment found that the footbridge was in good condition based on the threshold levels from the sensors [29].
The SHM at Pedro Gómez Bosque footbridge is constrained by the singularity and slenderness of the structure, keeping the aesthetic and cost restrictions [30]. These constraints have led Iban et al. [30] to design a low-cost vibration sensor based on triaxial MEMS accelerometer sensors embedded inside the handrail of the footbridge. The sensor circuit board used the ADXL327 MEMS accelerometer as the heart of the vibration measurements [30]. Iban et al. [30] validated the MEMS accelerometers with conventional piezoelectric accelerometers mounted in a portable analyzer. Data sampled at 200 Hz by CompactRIO 9076 was saved each hour for post-processing and to prevent measurement failures [30]. Based on the tests conducted on the footbridge, Iban et al. [30] reported that these low-cost sensors are a competitive alternative to traditional sensors.
Komarizadehasl et al. [31] presented a new “Low-cost adaptable reliable accelerometer” based on Raspberry Pi and Arduino called LARA. Based on the tests from the laboratory, LARA’s noise density is 51 g/√Hz and the acquisition rate is 333 Hz [31]. LARA uses Python software to schedule measurements and Network Time Protocol for data synchronization [31]. The test was conducted at the field, on a 14 m short-span footbridge in Barcelona [31]. The measured natural frequencies from LARA showed a maximum error of 1.28% compared to a commercial high-precision sensor, HI-INC [31].
Wired sensor systems can raise installation and maintenance costs for SHM, particularly in large civil infrastructure projects [32]. To overcome these challenges, Navabian and Beskhyroun [32] designed two series of WSNs for data collection from various structures. The initial WSN design cost around 70 NZD, the second WSN cost around 150 NZD, while a commercial acquisition system from Crossbow GP combined with a National Instruments cDAQ-9184 cost around 4000 NZD [32]. The tests performed on the Wellesley footbridge in New Zealand produced promising results [32]. The first WSN has a comparable measurement for large-amplitude vibrations with the commercial system, whilst the second WSN is highly sensitive and precise to measure very low-amplitude vibrations for accurate estimation of modal parameters for large and stiff structures [32].
In a study, Sabato [33] deployed a newly developed WSN utilizing a MEMS accelerometer to an arch footbridge for vibration analysis. The transmitter of the WSN comprised a SiFlex 1600SNA accelerometer, signal conditioner, a 2.4 GHz transceiver, a signal conditioner and a commercial acquisition board with a 24-bit accuracy link to a PC [33]. Comprehensive tests conducted at Streicker Bridge concluded that the system is reliable for acquiring data to detect damage yet comparable with traditional Integral Electronics PiezoElectric accelerometers [33].
Distributed Fiber Optic Sensor Embedded in Textile
Distributed fiber optic sensors (DFOS) are valuable SHM sensors and have evolved into a new method for continuous monitoring because they can detect environmental changes and respond to external stimuli, including mechanical and thermal variations [34,35]. However, the main drawbacks of this monitoring approach are the fiber’s fragility and the complexity of installation [35]. Established on Biondi et al. [36]’s previous work, in three different publications [34,35,37], a DFOS implanted in fabric was instrumented, offering the rigidity needed for installation in structures like pedestrian bridges and enabling constant measurements collection. In their studies, a 28m DFOS embedded in textiles was established on a footbridge to detect vibrations generated by pedestrians. Dynamic strain measurements were gathered and analyzed with Optical Frequency Domain Reflectometry [35]. Figure 8 presents two-year sensor responses compared. These three studies found that DFOS embedded in textiles is useful for SHM, enabling continuous, distributed sensing of strain measurements.
Large Area Electronic Sensor Sheet
Damage detection often requires direct sensing because anomalous behavior is typically close to the cracks. However, on-site sensor measurements are costly due to a dense array of individual sensors needing to be deployed [40]. A group of Princeton University researchers [40,41,42] proposed a novel sensor solution, Large Area Electronics combined with semiconductors for real-life applications. This sensing sheet, Figure 9b, consists of a concentrated group of resistive strain gauge sensing units printed on a flexible polymer sheet with CMOS integrated circuits for sensor control and readout [43]. These sensing sheets are glued with Araldite 2012 to the lower surface deck of the Streicker Bridge. The composite layers of the sensor are illustrated in Figure 9a. Based on Kumar et al. [40]’s report, the strain values from these sheets compared with the existing Fiber Bragg Grating (FBG) sensors matched closely, indicating that the sensing sheets are suitable for monitoring real-life structures. Tests conducted by Aygun et al. [42] revealed that the sensor is capable of detecting damage under the sensing area. Based on the conclusions from these three research, the experiments demonstrated that the sensor manufactured using cheaper and mass-produced printed circuit board technology is suitable and crucial for damage-detection systems.
Microwave Interferometer
Microwave remote sensing is one of the recent non-contact measurement techniques for deflection measurement in static and dynamic conditions on large structures [44]. In the research, Gentile and Marrongelli [44] proposed a microwave interferometer as a sensor unit. This sensor unit, a coherent radar emitting electromagnetic signals at 17.2 GHz midpoint frequency, is made from a sensor module, a power supply and a control PC [44]. Acceleration data of the footbridge was first collected with uni-axial WR 731A piezoelectric accelerometers, and then the data was compared with those collected from the microwave interferometer [44]. In the research conclusion, Gentile and Marrongelli [44] reported that the sensor provided precise and comprehensive data on displacement, velocity, and acceleration when influenced by the narrowband frequency content of running and walking-induced excitations.
Self Sensing Concrete
Current SHM is costly due to the wired connection, and WSN is unable to communicate inside concrete because of the concrete Faraday cage and battery replacement [26]. To overcome these issues, Gong et al. [26] proposed and developed a concrete with a built-in sensor using EcoCapsule. This EcoCapsule, a tiny low-cost piezoelectric energy harvesting sensor harvest its energy wirelessly and communicates by elastic mechanical waves using the Pulse Interval Encoding data coding scheme [26]. According to Gong et al. [26]’s report, a piezoelectric transducer (PZT) first injects a downlink elastic wave to a PZT in EcoCapsule embedded inside the concrete, a receiving PZT then converts the mechanical waves into electrical signals and collected by the piezoelectric backscatter system to power the logic and sensor circuitry. Gong et al. [26] reported overcoming obstacles in the energy generation for sensors embedded in concrete and data communication, and the results demonstrated a data transmission rate of 13 kbps and the ability to power the sensor from the range of 6 m.
Ultrasonic Sensor
In research by Terrien et al. [45], miniature sensors made with ultrasonic patches using active-guided-waves and passive-acoustic-emission methods were designed and tested. The ultrasonic patch selected was a 10mm radius, 200 μm thick piezoelectric disc [45]. The research found that these sensors can detect fiber breaks with low sensitivity compared to standard Acoustic Emission sensors [45].
Vision Sensor
The use of computer vision to monitor structural movements from video has reached wider endorsement for SHM [46]. Visual or image-based strategies are commonly employed to measure the active vibrations of bridges [47]. This section examines eight publications in Table 12 that develop sensing methods based on vision.
Using the Leica Geosystems MS50, a modern Image Assisted Total Station (IATS), Ehrhart and Lienhart [49] reported the development of a vision sensor suitable for measuring a structure’s absolute 3D coordinates and the eigenfrequency. Up to a few millimeters of accuracy was achieved, according to Ehrhart and Lienhart [49]. Data from the IATS experiments conducted on the Augarten footbridge compared to the reference measurements obtained using an HBM B12/200 accelerometer for the dominant frequencies matched almost perfectly [49].
In another research by the same researchers, Ehrhart and Lienhart [50] presented the IATS vision sensor for movement and deflection monitoring of structures. Ehrhart and Lienhart [50] reported that the sensor can detect movements with accuracy better than 0.2 mm up to a distance of 30 m from a single video frame. When averaging multiple video frames, accuracies of 0.05 mm are possible [50]. Ehrhart and Lienhart [50] also reported that the slow 10 fps rate is too low to transmit the video frames of the IATS to an external computer. But, the eigenfrequency of pedestrian bridges in the vertical direction is only 1.25 Hz ≤ ≤ 4.6 Hz [55]. Thus, the current IATS is adequate for measuring these natural frequencies, with future models anticipated to offer enhanced video frame rates. [50].
Using UAVs for sensor deployment was previously discussed. A new approach to using UAV cameras as a vision sensor proposed by Hoskere et al. [46] addresses several challenges related to structural modal analysis. Hoskere et al. [46] evaluated the proposed sensing method using a vibration table to excite a multistory model in the lab and validation on a real suspended footbridge. From Hoskere et al. [46]’s report, the laboratory experiments on multistory shear-building yielded trustworthy outcomes with errors in eigenfrequency below 0.5% and modal assurance criterion of more than 0.996. Further tests conducted by Hoskere et al. [46] on a real suspended footbridge in Illinois, produced MAC values above 0.925 and errors in the natural frequencies less than 1.6%. The outcomes highlight the effectiveness of the suggested image-based sensors for conducting structural dynamic assessment on real live structures [46].
With an inexpensive off-the-shelf camera, GoPro Hero 3-Black Edition, Brown et al. [48] introduced an innovative image and laser-based deflection sensor for monitoring the movements of structures from a remote location. Brown et al. [48]’s method differs from other visual-based techniques because the camera is placed at the subject structure. The illustration of Brown et al. [48]’s laboratory test setup is in Figure 10. Brown et al. [48] reported that under laboratory conditions, the sensor’s accuracy was about ±0.9 mm with 95% confidence from a range of 30.5 m. Additional testing on a pedestrian footbridge in real-world situations, including static and dynamic loading, showed that the prototype is cost-effective and functional [48].
In SHM, image-based techniques are typically designed for two-dimensional measurements of vibration displacement and crack detection [51]. It is hard to measure accurately the slight three-dimensional movements on real-life infrastructures [51]. Using two video cameras, Shao et al. [51] proposed a vision sensor without targets for tiny three-dimensional movement deflection aided with AI and movement amplification. The two video cameras formed a binocular system with an epipolar geometry-based approach, followed by applying a movement amplification formula to enhance the video and to make it easier to determine the slight three-dimensional deflection [51]. Shao et al. [51] reported that the vision sensor system precisely measured slight deflections and obtained the eigenfrequency in a three-dimensional environment.
Over the past decade, numerous studies employing vision-based approaches for displacement tracking have primarily depended on external processing of recorded video data [52]. In a study, V. Shajihan et al. [52] proposed a wireless SHM utilizing near-device computing to process video images and send measurements directly from local nodes. The sensor evaluates movements through target and non-target methods [52]. The equipment forming the sensor for this research was an OpenMV camera and wireless smart sensor, Xnode, developed by Fu et al. [56]. The operating system powering the Xnode is called FreeRTOS, whilst OpenMV uses Micropython [52]. Figure 11 illustrates the functional flow diagram of the Xnode. Several challenges discussed include sensing duration limitation, temporal aliasing, poor light situations, vision vibrations, and the waking up of the sensor due to insufficient power [52]. The accuracy reported by V. Shajihan et al. [52] achieved an average subpixel of 0.1555 pixels and 0.205 pixels. A MAC value of 99.31% attained from a field validation of the system on a footbridge demonstrated a strong connection with measurements in vibration analysis [52].
In the field of inspection for wood structures, Walker et al. [53] proposed an Eulerian-based virtual visual sensor (VVS) using a standard video camera to characterize structure vibrations. In the experiment, Walker et al. [53] recorded the transverse vibration of the subjected beams with the Metriguard 340 E-computer system to obtain the eigenfrequency and evaluate it in comparison with the outcomes of the VVS technique. A footbridge was employed to validate the visual sensors further [53]. The report concluded that the natural frequencies obtained using the VVS study were comparable to measurements from commercial E-computers [53].
Robotics has been extensively studied long before SHM technologies gained substantial attention [54]. It was not until recent years that robotic prototypes reached a level of maturity suitable for real SHM applications [54]. To date, robotic development for SHM has primarily concentrated on structural assessment compared to constant tracking [54]. These inspection robotic systems mostly use a camera mounted on a mobile robot or vehicle to capture images [54]. In the publication, Wang et al. [54] concluded that the recent advancements in depth-image (RGB-depth) sensor technology are anticipated to accelerate research in this area.
Wireless Mesh Network Sensor
Domaneschi et al. [57] proposed a dynamic wireless sensors network combined with an embedded fiber optic system that can complement each other and provide a wide range of information. The proposed solution consists of 14 synchronized WSNs fitted with a tri-axial accelerometer, a tri-axial gyroscope, a magnetometer, and a temperature sensor distributed on the Streicker Bridge [57]. According to Domaneschi et al. [57], the proposed WSN sets a new benchmark for inertial sensor performance with power consumption for each unity of about 400 µA, with the accelerometer noise performance 11% better, gyro noise performance 3% greater, and angular range of the compass over 4% improved when comparing with other sensors.
3.2.5. Sensor Fault Detection
Fault detection and identification are essential for reliable SHM systems but have received limited attention [58]. In the research, Al-Nasser et al. [58] shows a technique, the identification of combined sensor faults (ICSF) approach, aimed at identifying sensor faults using time-series data with a long short-term memory (LSTM) network. The ICSF approach was verified with acceleration measurements from a pedestrian footbridge. The outcomes reveal the performance of classification models in detecting various combined sensor malfunctions. However, classification models have faced challenges in detecting outliers due to the imbalanced training data resulting from their irregular occurrence in continuous signals.
3.2.6. Sensor Validation
Sensor validation is essential to ensure their accuracy, reliability, and conformity to specifications. Research on sensor validation aims to confirm the accuracy of sensors by comparing their measurements with those obtained through more conventional methods.
Low-Cost Accelerometer
Using a dynamic sensor (HI-INC) obtained commercially, Delgado et al. [59] presents an experiment to validate LARA. The results demonstrated it has a lower noise density than the MPU9250 accelerometer and a maximum error of 1.28% compared to the HI-INC commercial dynamic sensor.
Microwave Interferometer
The IBIS-FS is a recently developed interferometric radar system for SHM [60]. In recent research, Sofi et al. [60] focuses on using IBIS-FS for the detection of movements in a footbridge’s eigenfrequency and validating the interferometric measurements with data acquired by conventional accelerometer measurements. The research produced an outcome showing significant regularity among the two datasets, concluding that interferometric radar technology can potentially used for SHM and assessment [60].
3.3. Data Acquisition
Data acquisition implicates the acquisition of analog signals and the transforming of data into digital signals for storage and analysis [61]. In the SHM of footbridges, there are various methods of data acquisition. This section will look through five subcategories of data acquisition topics identified from the collection of literature.
3.3.1. Crowd Sensing
Crowd sensing, a technique of using devices equipped with sensing capabilities such as acceleration and global positioning systems like smartphones, smartwatches, and autonomous vehicles, has become common [62]. Unlike traditional sensor systems, crowd sensing features nonstationary, mobile, and distributed sensor network components [63]. Mobile sensors carried by pedestrians or moving vehicles pose challenges for measuring structural vibrations, as the sensor data include both the structure’s vibrations and the biomechanical characteristics of the mobile sensors [63].
Drive-By
Indirect SHM has recently explored drive-by tracking of footbridges using the vibration analysis of vehicles mounted with sensing equipment [64].
An innovative drive-by indirect SHM strategy, utilizing data collected from smartphones temporarily mounted on bicycles, was proposed by Quqa et al. [64]. This proposed strategy is an initial demonstration that displays the potential of determining high modal density that could be particularly useful to localize structural damage [64].
Another drive-by study for SHM conducted by Peng et al. [65] presents a portable crowdsourced data collection technique for the identification of tightly spaced spatial sampling footbridge modal density with the use of a scant mobile-based dataset. Experiments conducted with remote-controlled vehicle models demonstrated that Peng et al. [65]’s method could be integrated with modal shape-driven deterioration characteristics to identify defects in the structure [65].
Smartphones
The feasibility study of smartphone accelerometers as SHM sensors conducted by Ozer and Feng [66] on two pedestrian bridges produced satisfactory results. The results demonstrate that smartphone sensors can effectively measure structural vibrations and identify modal frequencies compared with high-quality reference sensors [66]. Ozer and Feng [66] also noted that, based on the experiment results, the precision of the gauged magnitude is significantly influenced by the sensor’s position. Shifting the sensor location from mid-span to one-third span reduces the peak vertical acceleration error from 110% to 32.7% [66].
Crowdsensing in SHM incorporates device mobility, significantly altering outcomes caused by ambiguities in both spatial and temporal domains [67]. To address these challenges, Ozer and Feng [67] presented a vibration modes determination design with parameters derived from multi-stream measurements collected independently through smartphone-based WSNs. When compared with a regular monitoring system, the suggested technique demonstrated that WSN data infrequently occurs across space and time can reliably estimate modal characteristics.
A further challenge in crowd sensing with mobile devices is isolating mechanical influence on the body from the dataset to retrieve the deflections and modal parameters [63]. The study by Ozer and Feng [63] on two primary human movements, standing and walking, aimed to evaluate walking load and determine vibrational characteristics, marking a pivotal step toward transitioning from traditional SHM systems to community-collected, mobile phone SHM methodologies.
Accurately extracting real-time data exceeding the natural frequencies of a structure remains an ongoing problem and necessitates addressing the issue from a measurement viewpoint [62]. In the investigation by Iacussi et al. [62], which investigated the effect of device synchronization on reconstructing structural deflection shapes, Iacussi et al. [62] noted that for infrastructures with a low resonant frequency, typical of civil structures, the errors caused by synchronization are less critical. However, these errors become significant in the high-frequency range aiming at reconstructing the operational deflection shapes [62].
3.3.2. Design Validation
In a study, Miskiewicz et al. [68] installed an SHM to perform data acquisition and design validation on a footbridge in Radom, which failed acceptance tests. The proposed remediation was tuned mass dampers (TMD). However, the current footbridge structural design must be validated to match theoretical calculations before starting remediation work [68]. The experiment collected data on wind power, movement, the temperature of the air, pressure in the atmosphere, acceleration, and angle over six weeks [68]. The experiments found that the eigenfrequency of the footbridge from the measurement period of 2.6 Hz on the cable and 1.6 Hz on the pedestrian bridge deck corresponds to the outcome of hypothetical computations [68].
3.3.3. Internal Features Mapping and Material Characterization
Ground penetrating radar (GPR) is a highly effective instrument for structural investigations, offering valuable insights ranging from contaminant distribution to crack localization [69]. In two different studies, Morris et al. [69] and Morris et al. [70] used GPR to map the footbridge deck’s inner element and determine the dissimilarities of the concrete in two building stages. Alongside standard processing and filtering techniques, Morris et al. [70] analyzed additional data attributes to quantify the differences in concrete strength between the construction phases [70]. Morris et al. [69] reported that information from the study could validate visual interpretations, detect construction anomalies, and provide essential details for future decision-making and SHM. Attribute analyses for evaluating material properties after validation will facilitate more informed, comprehensive, and unobtrusive SHM [70].
3.3.4. Method Development
Method development involves creating a reliable approach, qualification assessment, and validation to ensure the method developed complies with the requirements for its intended purpose. This section explores publications dedicated to data acquisition method development in SHM.
E-Waste and Recycle Materials
The rapidly growing demand for electronic gadgets is leading to a significant increase in electronic waste (e-waste) [71]. Peralta and Smarsly [71] proposed using sustainable wireless sensors from e-waste for the SHM of a footbridge. With the help of a traditional WSN as a comparison, the eco-friendly SHM method was verified on a footbridge [71]. The research reported that the experimental results showed that reclaimed electronic waste parts are feasible for implementing eco-friendly SHM systems [71].
Image Assisted Total Station
Ehrhart and Lienhart [72] proposed using a commercial image-assisted total station (IATS) for motion and deflection data acquisition for SHM. The research demonstrated that the IATS can determine deflections to a resolution of 0.1 mgon performed under laboratory investigations [72]. During an on-site trial on a life-size pedestrian bridge, the main frequencies of the pedestrian bridge’s vibration were precisely determined from the IATS data. However, the primary limitation of the acquisition with IATS is its restricted frame rate of 10 fps.
Network-Based Real-Time Kinematic GNSS
Yu et al. [73] proposed a GNSS system with network-based real-time kinematic (NRTK) for tracking bridge vibrations. According to Yu et al. [73], this data acquisition method is the first of its kind. The laboratory and field experiments were validated using traditional real-time kinematic (RTK)-GNSS in conjunction with accelerometer methods on the Wilford suspension footbridge [73]. Figure 12 reveals three dynamic displacement time series derived from the experimental data. Data from the NRTK-GNSS successfully identified the active vibration and frequencies on the footbridge with satisfactory accuracy [73].
3.3.5. Structural Behavior
The numerical simulations performed by Miśkiewicz et al. [74] on a single-cable suspension footbridge before building suggested that there may be structural issues with the design. Active load test on the real bridge produced up to 4.5 m/s2 accelerations with the risk of structural damage and user discomfort [74]. Two TMD dampers were installed on the footbridge deck together with displacement sensors and accelerometers to facilitate repairs and monitor structural behavior. Dynamic load tests conducted after the repair showed positive results [74].
3.4. Data Processing and Analysis
Data processing and analysis involves transforming raw data into meaningful information through structured steps such as collecting, organizing, analyzing, and interpreting data using manual or automatic processes to generate valuable insights for decision making. The data processing and analysis category represents the most research area among the publications reviewed, accounting for 50.3% of the included studies.
3.4.1. Attribute Analysis
Morris et al. [75] conducted a GPR survey on Streicker Bridge as an initial study of the usefulness of characteristic examination methods. Common processing strategies and attribute evaluation approaches were used to analyze GPR echoes from two construction phases, identifying construction elements and comparing attribute signatures of concretes with different strengths [75]. The studies on the bridge highlight the possibility of using the attribute analysis technique for substance characterization, particularly to complement different SHM and non-destructive evaluation methods [75].
3.4.2. Concrete Behavior Prediction
Predicting the concrete extended behavior is challenging due to the probabilistic character of its flow-related properties [76]. To address these hurdles, Pereira and Glisic [76] suggested a hybrid method combining probabilistic neural networks and engineering code models to forecast the extended behavior of concrete infrastructures. The method allows the identification of abnormal structure behaviors with progressive and abrupt occurrences precisely forecasted years beyond the training range [76].
3.4.3. Damping Estimation
Bhowmik et al. [77] highlights that instantaneous evaluation of damping ratios in civil infrastructures is crucial for continuous health monitoring. Bhowmik et al. [77] demonstrated an automated enhanced frequency domain decomposition (AE-FDD) technique in assessing real-time damping on Daly’s “Shaky” bridge, a lively suspension bridge in Ireland. The AE-FDD method offers precise calculations of modal damping, demonstrating the durability and effectiveness for practical use cases [77].
3.4.4. Data Correction
Data synchronization discrepancies, sampling jitter, and sensor anomalies leading to low-fidelity measurement data and data loss are challenging fields that Ngeljaratan et al. [78], Dragos et al. [79], Ramadan and Ozer [80], and Zhang et al. [81] focused on researching.
Ngeljaratan et al. [78] have successfully demonstrated compressive sensing techniques can improve target-tracking signals when encountering data loss in the research.
Dragos et al. [79] reported implementing and validating cross-spectrum synchronization on a WSN and wired SHM system in the lab and on a footbridge. The cross-spectrum synchronization method could complement traditional clock synchronization protocols, improving the accuracy of SHM outcomes from the verification experiments [79].
Ramadan and Ozer [80] analyzed the statistical attributes of mobile phone measurement intervals using kernel distribution and implemented a restoration solution to address the data timing fluctuations issue. The results demonstrate the success of restoration of smartphone measurements, effectively mitigating timing deviation errors and significantly enhancing detection precision [80].
Zhang et al. [81] presents an automated limited-sample classification technique for detecting sensor oddities using a restricted number of labeled specimens. The top discriminative time series subsequence is extracted from the datasets to locate the prominent unnatural features [81]. The method, when used in measurements gathered from two different pedestrian bridges, demonstrated that it was able to identify unknown oddities using a small number of data [81].
3.4.5. Design Validation
A SHM system consisting of strain, vibrations, and displacement monitoring was installed on a footbridge for foot and bicycle traffic to ensure presumed capability and to monitor time-dependent material property changes [82]. The data collected, processed and analyzed to validate the FEM formed during the invention process [82]. Strain monitoring results revealed a substantial influence of temperature variations on the acquired data while not observing notable global changes in displacements [82]. Additionally, motion tracking of the bridge indicated no alterations in the motion-related properties [82].
3.4.6. Displacement Analysis/Estimation
The Leica Nova MS50 MultiStation, an image-assisted total station, is potentially utilized as a modern geodetic measurement system capable of precise vibration and displacement analysis [83]. The outcome of the experiments demonstrates that the Leica MS50 is capable of precisely estimating the deflection and movement characteristics under 5 Hz [83].
In another displacement analysis research, Xu et al. [84] presented an economical visual recognition method for multiple location vibration acquisition, leveraging a regular consumer camera for video recording and a specifically designed method for motion picture analysis. The research validated the technique on a pedestrian bridge by measuring the structural deformation of the deck and cable vibrations caused by human load [84]. The analysis results of both measured parameters allow for precise prediction of the eigenfrequencies, and the results can be utilized to study divergences in eigenfrequencies under different human loading conditions [84].
Fradelos et al. [47] suggested that a simplified algorithm to process standard video imagery can revamp the two-dimensional movement dynamic of elastic footbridges. Fradelos et al. [47] applied this technique to a timber bridge with managed forced vibrations, where the footbridge was vertically rigid but exhibited significant lateral flexibility due to accumulated damage. Figure 13 displays the chronological data of calculated deformations for the reference positions. The research produced precise predictions of the eigenfrequency and the attenuation of the footbridge and was matched with predictions extracted by using other sensing devices [47].
Yu et al. [85] introduced a novel hybrid device, the “smartstation”, for monitoring displacements of footbridges. Full-scale investigations were carried out on a footbridge to assess the method’s viability using the composite instrument [85]. Yu et al. [85] reported that in experiments within the range of 5.7–10.0 mm and 1.3–2.5 mm, the technique precisely recognized almost stationary and active vibrations.
Ma et al. [86] proposed a technique for deflection prediction for structures that combines datasets from a frequency-modulated continuous-wave millimeter wave radar (FMCW) and an accelerometer. Figure 14 shows the overview of the design and flowchart. The deflection was predicted on-the-fly using the radar data with a phase-unwrapping algorithm with the aid of an accelerometer if the radar wavelength is smaller than the structure deflection [86]. The proposed technique produced an enhanced deflection estimation precisely [86].
3.4.7. Dynamic Assessment/Monitoring/Response
He et al. [87] concentrated on vibration recognition and FEM of a footbridge in the investigation. In between ranges of frequencies of 3.59 to 14.92 Hz, He et al. [87] discovered several consistent modes linked with dominant bending and twisting distortions using a highly synchronous tri-axial WSN, which later led to the design of four distinct FEMs [87]. According to He et al. [87], the conclusive FEM is suitable as a reference for continuous tracking and provides recommendations to scholars for analyzing and modeling this type of footbridge.
Huang et al. [88]’s research focused on OMA on a stress-ribbon footbridge. Huang et al. [88] determined the bond failure of the hollow steel profiles filled with concrete and the concrete elasticity modulus using ultrasonic instruments and rebound hammers, respectively. By comparing the identified FE model with non-destructive testing methods, Huang et al. [88] reported that the non-destructive testing methods provide crucial supplementary information to the SHM.
Colmenares et al. [89]’s dynamic assessment of the KTH Royal Institute of Technology’s concrete footbridge using the FEM and three-dimensional solid FEM in autumn and winter weather identified that the weather conditions significantly influence differences in eigenfrequencies of the studied footbridge. Colmenares et al. [89] reported that the findings will help minimize the ambiguities linked with an FEM.
Li et al. [90] presented a technique to determine a footbridge’s frequencies with pedestrians by analyzing the vibrations of a shared scooter. An innovative FEM consisting of four independent directions of motion was presented to simulate a rider and a scooter. The study highlighted several additional engineering factors between the vehicle and the bridge, which warrant further investigation.
The investigation of the patterns of vibration caused by human motion in footbridges with elevated vibrational frequencies and the assessment of the viability in a Tuned Liquid Damper (TLD) to dampen vibrations was conducted by Lu et al. [91]. Using a simple pure water TLD to test and analyze the feasibility of vibration control, Lu et al. [91] identified that the TLD decreases the bridge’s primary frequency and raises the attenuation ratio which provides relevant data for SHM and vibration control for similar structures under human-induced vibration.
Bassoli et al. [92] aimed to characterize the time-dependent response of a steel curved pedestrian bridge for ambient vibrations and pedestrian-induced loads while assessing the impact of temperature variations on structural modal properties using an SHM established on MEMS sensors. The structural motion-related properties are determined with the use of the classical Enhanced Frequency Domain Decomposition (EFDD) method [92]. The experiments found that the value of the vertical accelerations induced by wind reached about 3–5 mg while a crowd of walkers produced vertical accelerations higher than 100 mg as illustrated in Figure 15 [92]. Bassoli et al. [92] reported that a model updating procedure is needed to match the experimental results. The numerical dynamic investigation was conducted after refining the FEM to reproduce pedestrian dynamic loads [92].
Cunha et al. [93] described two dynamic monitoring systems installed on lively footbridges to characterize their dynamic behavior and response, with the potential to detect damage at early stages. By using LabVIEW software, Cunha et al. [93] developed a Continuous Structural Modal Identification framework for tracking of deviation of modal characteristics in both footbridges. The experiments presented results from the automated data analysis of the monitored footbridges, highlighting the possibility of constant vibration tracking [93].
Moutinho et al. [7]’s research work involved the numerical and experimental dynamic analysis of a very flexible pedestrian bridge in Porto, where it received a high level of vibrations reported by the users. The research begins with the description of the bridge and identification of its dynamic properties, followed by experimental and numerical simulations of various pedestrian loading scenarios exciting the bridge [7]. Using a dynamic monitoring system to characterize vibration levels, the results revealed that a single pedestrian walking under resonance conditions induced maximum acceleration on the footbridge exceeding 0.5 m/s2, which approaches the limit specified in several codes [7].
Bai et al. [94] presented an innovative framework to track bridge behavior, employing distinct camera locations and several image processing methods to achieve economical yet highly efficient results. The framework incorporates a deep-learning-based data processing and analysis technique, verified through an optical motion tracking technique for movement monitoring [94]. The framework incorporates a deep-learning-based data processing and analysis approach that demonstrated its accuracy in capturing motions in the laboratory and real-life footbridges [94].
A non-invasive, low-cost, vision-based technique proposed by Ponsi et al. [95] using image-based techniques for the vibrations tracking of footbridge demonstrated satisfying results. However, Ponsi et al. [95] also concluded that there are real conditions such as camera shakes, lighting, field of view and camera synchronization drawbacks that could potentially affect the results are areas for future studies.
Gamino et al. [96] employed a computational modeling approach incorporating a soil-structure interaction methodology to represent the vibration characteristics of a timber-concrete pedestrian bridge. The monitoring conducted with ground-based radar interferometry and video motion amplification highlights the significance of incorporating the soil-structure interaction approach to accurately represent the bridge’s dynamic behavior and develop reliable models for assessing its structural responses.
In the research, Ponsi et al. [97] investigated vision-based techniques based on target tracking for dynamic monitoring and identification of a steel footbridge. The vibration outcomes from the image-based technique, when corresponding with accelerometer data, show satisfactory agreement in time-space and frequency-space, resulting in the identification of three eigenmodes of the footbridge, demonstrating the possibility of the image-based approaches.
In a study of a pedestrian bridge by Miskiewicz et al. [98] that includes computational modeling, on-site vibrational loading experiments and an SHM system, the pedestrian bridge was found to possibly have structural dynamics problems based on the numerical simulation results. During dynamic loading, the deck exhibited movements with accelerations reaching as high as 4.5 m/s2, consistent with the findings of the computational simulations [98]. The repair project involved installing TMDs on the deck’s external girders at its mid-span, a design validated through numerical simulations and analysis [98]. The result before TMDs and after TMDs is illustrated in Figure 16. Repeated dynamic tests performed on the bridge following the repair work confirmed the positive outcome [98].
In a footbridge response characterization study, Kromanis [99] presents the findings on surveying distortions of a pedestrian bridge with loading from cyclists using an altered GoPro camera equipped with a lens that is zoomable. Kromanis [99] reported employing techniques such as visual data processing, de-noising recorded data, interpretation of vertical displacements, and the derivation of influence lines. The method can detect stationary and active motions of footbridges, and vertical deflections calculation in sub-millimeter accuracy from the camera recording [99].
Kromanis and Elias [100] used the templating recognition method to estimate the subject’s pixel motion, from which precise reference deflections can be obtained. These displacements are utilized to determine the frequencies of the movements and generate the shape of the first vertical oscillation mode [100]. In conclusion, Kromanis and Elias [100] reported that despite being able to gauge high-mode vibrations, determining oscillation mode shape is challenging when using 1/10 pixel resolution data.
3.4.8. Heat Transfer Analysis
In a heat transfer analysis study, Xia et al. [101] designed an integrated method that combines thermal conduction and structure investigations to estimate the thermal profile and related structural reactions by integrating field monitoring data. Figure 17 illustrates the sensor placement and measurement collection units. The computed thermal data was used as input for the FEM model to determine the footbridge’s thermally induced reactions [101]. The computed and surveyed thermal reactions demonstrate strong alignment. The newly invented integrated method allows effective analysis of the temperature responses of footbridges automatically [101].
3.4.9. Image Processing Analysis
Do Cabo et al. [102] presented a phase-based motion extraction (PME) technique, that was opposed to a hybrid approach combining template matching and particle filter for amplification and derivation of frequency data. The presented techniques demonstrated significant potential for extracting real-time data in the absence of markers on the subject at the same time, showing precision compared to PME prediction and off-the-shelf software [102].
3.4.10. Load Prediction
Identifying walking patterns and predicting pedestrian loads on a footbridge is challenging due to the complex ground reaction forces generated by human-induced vibrations [103]. Sadhu et al. [103] performed an investigation demonstrating a wavelet-based time-frequency decomposition technique for extracting pedestrian walking patterns using video cameras and vibration sensors. Full-scale investigations and experiments suggested that the technique has accuracy in identifying walking patterns [103].
Hassoun et al. [104] presented a novel technique for continuous crowd load prediction on pedestrian bridges, focusing on high-density crowds using FBG fiber optic sensors and machine learning (ML) to generate prediction models. Hassoun et al. [104] investigated different types of bridge loading scenarios to predict the weight of individuals and groups of people and the continuous flow of people’s activities in the laboratory and on a scaled bridge model [104]. Figure 18 reveals the result of the groups of people for fast and slow activities. The findings highlighted the capability to estimate the total load weight with a mean error of 10 kg and determine activity speed with over 90% accuracy [104].
Mustapha et al. [105], focusing on pedestrian load prediction, introduced effective methods for estimating crowd flow and bridge loads by leveraging cutting-edge AI and measurements collected from structure sensing devices and body-worn sensors. The AI approaches, including Convolutional Neural Networks (CNN) and Support Vector Machines (SVM), were initially used in individual sensing sources, such as FBGs and accelerometers, and later extended to enable multi-sensor measurements integration [105]. The results demonstrate the monitoring solution’s effectiveness, achieving 98% precision for one movement type speed identification, 91% for multiple movement type speed and loading assessment, and a minimum 9% error in load behavior prediction [105].
3.4.11. Method Validation
Liang et al. [106] presented a technique utilizing 3D imaging with laser for assessing mechanical performance. Initially, the study derives a reverse analysis approach alongside a conceptual model for evaluating the physical characteristics of pedestrian bridges [106]. The study develops a point-based model of the footbridge to extract alignment and key parameters, deduces forces on key components using back analysis, and verifies the precision of the outcomes by corresponding them with the inner force measurements [106]. When compared to traditional single-point measurements, the results present an alternative and efficient approach for the assessment of the mechanical behavior in complex bridges [106].
To overcome the limitations of wired solutions and costly equipment, Casciati et al. [107] proposed an innovative method using a measurement integration method based on Kalman filtering, integrating reactions from a GPS and an acceleration sensor. GPS accuracy is improved by utilizing satellite refinements from a single base reference using the Real Time Kinematics method [107]. The method was validated on the “Tesa” cable-stayed pedestrian bridge [107].
The study by Ngeljaratan and Moustafa [108] explored marker tracing digital image correlation in the role of contactless data acquisition technique. The method was validated through a large-scale laboratory application, monitoring the reaction of a footbridge subjected to two-directional earthquake movements [108]. The digital image correlation method validation successfully captured the footbridge’s deck rotaional movement and deflection caused by varying grades of seismic severity [108].
Barros and Paiva [109] proposed a non-intrusive radar interferometry method for SHM and validated the method on four different bridges in Portugal. The radar sensor used in their study, the IDS IBIS-FS system, is an industrial-grade microwave interferometer that includes a sensing device, a personal computer, and an electrical energy source [109]. The method was validated for natural and excitation frequencies to produce high accuracy, portability, and ease of use of the non-intrusive radar interferometry technique [109].
A temperature compensation methodology developed by Weil et al. [110] has laid the foundation for Particle Filter-based temperature compensation of FBG strain data. This Particle Filter-based implementation, compared with the Kalman Filter, was validated to have improved results by modeling the strain events as an asymmetric Weibull Distribution.
In Ponsi et al. [111]’s research, the method of fusing Global Navigation Satellite Systems (GNSS) data with accelerometers was validated on a steel footbridge. The data-fusion outcomes demonstrate that the technique is able to mitigate the constraints of single sensing devices and deliver better displacement estimates precisely.
3.4.12. Modal Identification
Abasi and Sadhu [112]’s study evaluated and compared Time-Varying Filtering Empirical Mode Decomposition (TVF-EMD) and Second-Order Blind Identification (SOBI) on a large-scale footbridge and a wind rotor blade with diverse dynamic characteristics. Abasi and Sadhu [112] reported that the outcomes indicate that TVF-EMD achieves higher precision in modal analysis than SOBI. Nevertheless, SOBI offers greater computational efficiency over TVF-EMD when the sensor number is equal to or larger than the modes of interest’s number.
Brownjohn et al. [113] detailed using wireless inertial measurement units as sensors, developed for biomechanical movement analysis, in the vibration assessment experiment of a footbridge. Brownjohn et al. [113] used the sensors to determine the modal frequencies and damping parameters in an output-only dynamic assessment and to measure the inertial responses of a jumping person, correlating them with the deck’s dynamics. The mode-specific mass predictions of the footbridge, when compared with those acquired from a hammer with a force transducer and specified mass configurations, demonstrate uniformity throughout the investigation results [113].
Górski et al. [114] performed an investigation on monitoring the vertical direction of a glass fiber-reinforced polymer (GFRP) footbridge. The research included mode shape recognition and vibration-based experiments with several specific vibrational inputs based on ambient vibration tests [114]. Górski et al. [114] identified seven eigenfrequencies reaching as high as 21 Hz, associated vibration modes, and energy dissipation ratios of the pedestrian bridge. Figure 19 illustrates the results from the experiments ranging up to 21 Hz. The outcomes delivered valuable information for the SHM of identical structures subjected to comparable loading situations [114].
Jiang et al. [115] presented an innovative approach for dynamic response reconstruction of structure and virtual sensing in SHM, utilising a sequential modeling method. This method leverages space-time dependence in sequential measurements to enhance data transfer, significantly improving reconstruction performance and demonstrated to be effective [115].
To fully utilize the potential of decentralized implementation, Sadhu et al. [116] developed new algorithms that leverage sparsity concepts and wavelet transformation under the blind source separation (BSS) approach. The identification challenge is framed under the underspecified BSS framework, applying measurement transformations to the time-frequency domain to achieve a compressed approximation [116]. The proposed methods experimented on a live infrastructure using WSN in a decentralized deployment successfully recognized the structure modes of the subjected bridge with only very little data in a decentralized manner [116].
To overcome challenges in contactless image-based dynamic mode analysis techniques for SHM, Banerjee and Saravanan [117] proposed a blind system determination method to retrieve eigenfrequencies and modal shapes from the acquired video recordings. Figure 20 illustrates an outline of the proposed methods. The computation inaccuracies were verified below 1%, making it highly applicable and dependable for SHM [117].
Domaneschi et al. [118]’s study focused on the vibration mode characterization of a pedestrian bridge by output-based method. The investigation utilizes measurements from an extended-gauge fiber optic stress sensing device implanted throughout the pedestrian bridge, with the spectral power distribution of the dispersed extended-gauge vibration-induced curvature being employed [118]. Comparison with previous studies in the literature validates this presented technique, although observing narrowly reduced frequency measurements [118].
To leverage advancements in SHM and explore their use case for lightweight footbridges, Wang et al. [119] presented a case study of an aluminium alloy footbridge in a high pedestrian activity area. Acceleration time series at specified structural joints were collected for OMA, along with recordings of pedestrian traffic on the structure [119]. The prototype study concluded that the applied methods are effective for data processing, modal identification, and model updating in lightweight footbridges [119].
In Ozer [120]’s research, a medium-term vibration data set from smartphones handheld on a bridge performing different tasks was presented. The result demonstrated that the dataset observations are discontinuous but may still provide insights into the long-term impacts that could have caused fluctuations in the bridge’s natural frequency.
Feng et al. [121] proposed a modal identification utilizing the drive-by method using smartphones and application-specific accelerometers installed on electric scooters and bicycles. Validation on a pedestrian bridge showed that smartphone sensors can deliver precise output comparable to application-specific accelerometers. The footbridge’s eigenfrequencies were extracted from stationary and moving scooters and bicycles using variational mode decomposition and filtering techniques.
Hacıefendioğlu et al. [122] introduces an advanced model identification method by combining multiple visual recognition methods. The result shows substantial improvements in distant and contactless SHM.
Qiao et al. [123]’s research aims to investigate a lightweight synchronization concept using two long-range (LoRa) wireless sensing modules for mode shape and frequency extraction. The test results demonstrated that the synchronized LoRa nodes can rapidly and accurately obtain an average synchronization precision of the principal mode of the beam vibrations of 4.5 ms.
3.4.13. Model Development/Prediction/Validation
He et al. [124] performed an OMA on a footbridge kept on a butterfly arch. A precisely timed three-axis WSN was deployed for the capturing of the footbridge’s environmental movement responses induced by the breeze and light pedestrian traffic [124]. The experiment results obtained through the subspace-based system identification technique revealed that the experimentally obtained mode shapes exhibited in-phase and out-of-phase vertical deformation, along with some torsional dynamic modes [124]. He et al. [124] reported that a model refined by a sensitivity-based method may operate as a reference instance for extended tracking of the footbridge’s health.
Venuti et al. [125] presented and validated 2 separate FEM of the Streicker pedestrian bridge. Venuti et al. [125] first developed the 3D beam-based model based on the detailed drawings. Thereafter, an enhanced discretization of the footbridge deck was implemented using shell-based elements [125]. The footbridge FEM was verified using pre-recorded SHM measurements from stationary and active experiments [125]. The results indicate that the S model surpasses the B model in accurately capturing the bridge’s stationary and active behavior, particularly between the primary span and supports, and is more satisfactory in representing the geospatial spread of human loads [125].
Catbas et al. [126] investigated digital recording and model design for infrastructures lacking reference materials by using data from LiDAR technology to acquire the datasets of a footbridge. The datasets were analyzed for the generation of geometrical and structure divisions. The investigation demonstrated that the datasets generated return favorable outcomes specifically for footbridges lacking reference materials.
Li et al. [127] proposed an improved hierarchical Bayesian modeling framework to address issues with high computational costs in the hierarchical Bayesian modeling framework. The results verified via the investigation of a footbridge demonstrate that the framework provides a higher precision assessment of uncertainty in structural parameters and shows increased computational performance analogized with the hierarchical Bayesian modeling framework.
While FEM refinement methods were widely investigated recently, investigations established on deterministic techniques still dominate existing academic work, failing to take into consideration the unpredictable consequences in the process of model refinement [128]. Yin [128] proposed a functional approach for FEM refinement and estimation using Bayesian regularization with insufficient datasets. The approach was anticipated to increase the performance of model refinement and estimation [128].
Ao et al. [129] proposed a low-cost sensor, highlighted the impact of input parameters, proposed a modified Principal Component Analysis (PCA) model optimization method and explored data models developed with limited data set. Ao et al. [129] reported successfully detecting damage using the single-class SVM AI method that leverages results from data modeling.
Castle Cornet Bridge is a nine-span reinforced concrete bridge subjected to corrosion and deterioration described in the study of Lutton et al. [130]. FEA modeling of the footbridge was conducted before the load test, with Structural Health Monitoring (SHM) implemented using the FBG sensing device to acquire the bridge’s structural reaction during the test [130]. Lutton et al. [130] used the SHM data obtained from FBG sensors in comparison to the FEA model for the validation of the theoretical model [130].
3.4.14. Multivariate Statistical Analysis
Bjorngrim et al. [131]’s study involved evaluating the suitability of PCA for data obtained from SHM system. Bjorngrim et al. [131] employed principal component analysis and partial least squares modeling to gain more interpretation of the complicated relations of bridge responses and climatic influences. Bjorngrim et al. [131] reported that the results demonstrated PCA provides a comprehensive overview of the collected data, while partial least squares modeling revealed that wind explains the bridge movements.
3.4.15. Operational Modal Analysis
Komarizadehasl et al. [132] sought to overcome common challenges in low-cost SHM solutions by installing 4 LARAs as illustrated in Figure 21, on a short-span footbridge to conduct experiments. The study focused on automating data acquisition and management, as well as creating a digital model in SAP2000 established on available illustrations and physical properties [132]. The bridge’s OMA was conducted using the FDD and Covariance Stochastic Subspace Identification methods [132]. Komarizadehasl et al. [132] documented that the footbridge’s eigenfrequencies measured using an ultra-precise sensor, the HI-INC, closely matched those obtained using LARA.
Sun et al. [133] further explored the traditional transmissibility-based operational modal analysis technique, which uses data with various load situations, and the power spectral density transmissibility technique, which relies on multi-reference outputs subjected to a one-load situation. Sun et al. [133] proposed that addressing uncertainties is crucial for enhancing the reliability of recognition outcomes and improving algorithm precision. Limited research has addressed the unpredictability effects of the power spectral density transmissibility technique, highlighting the need for future unpredictability investigation in this area [133].
In the study, Fontan and Guerineau [134] utilized new and standard sensors on a 70-year-old retired concreted footbridge to conduct OMA. The measurements were assessed and contrasted with several formulas, including SSI and EFDD, and these formulas were subsequently used to fine-tune an FEM of the footbridge with the MAC matrix [134]. Fontan and Guerineau [134] reported that incorporating dynamic properties enhances the understanding of the structure and helps prevent incorrect conclusions regarding the use of the aging bridge.
3.4.16. Prestress Losses Analysis
Prestressed concrete has become increasingly popular in bridge construction due to its superior qualities as a building material [135,136]. Prestress losses affecting the load limit, strength, and strain capacity of prestressed concrete have been widely investigated by researchers [137,138,139], with particular focus on those caused by corrosion. Prestress losses along the structure are one of the key markers of the structure’s condition and performance of the prestressed concrete structures [135,136]. Abdel-Jaber and Glisic [135] introduced a novel monitoring method for prestress losses using long-gauge fiber optic sensors implanted inside the concrete, including treatment of shifting environmental conditions, for example, thermal and logging the strain and temperature on site. Abdel-Jaber and Glisic [135] reported that the measured prestress losses closely match the long-term design values provided by the bridge designer, with only minor differences. Prestress losses can be predicted with better results when combining a sensor-based SHM system with an ML-based framework, demonstrated by Calò et al. [140] on a scaled prestressed reinforced concrete box-girder bridge.
3.4.17. Rheological Deformation Prediction
An inaccurate examination of extended rheological characteristics can compromise functional performance, negatively impact pre-tensioning forces and lead to expensive repairs [141]. Pereira and Glisic [76] suggested an integrated approach combining probabilistic AI and structural code-based models to forecast the extended behavior of concrete infrastructures. The method demonstrated its capability to accurately predict both the rheological strain and the total strain for up to one and a half years beyond the training interval [141].
3.4.18. Static and Dynamic Analysis
Desbazeille et al. [142] suggested that static and dynamic damage analysis can be carried out using the identical equipment, consisting solely of MEMS accelerometers. Desbazeille et al. [142] demonstrated the application of both analyses on a live pedestrian bridge. Desbazeille et al. [142] compared stationary deformations and modal flexibility-derived displacements under functional conditions, with human loading and thermal variations, as well as with artificially induced damage. Desbazeille et al. [142] reported that stationary and dynamic analysis can complement each other, providing additional information to improve the reliability and precision of damage recognition.
3.4.19. Strain Prediction
Oh et al. [143] proposed a CNN data projection technique for extended stress data acquisition in concrete buildings, leveraging the close relationship between environmental thermal effect and structural reaction. Ambient temperature and related time data are used concurrently as the information for the CNN to capture the progressively extended behavior of a concrete building [143]. In one of the findings, Oh et al. [143] reported that, when using extended periods in the training’s data, the CNN could provide better accurate estimates of the structural responses [143].
3.4.20. Structural Evaluation/Identification/Material and Response Prediction
Górski et al. [144] presented a study on the changes in modal parameters and rigidity of the Fiberline Bridge, a cable-stayed structure constructed completely with Glass fiber-reinforced polymer hybrid, which had been in use for 20 years and exposed to extreme environments. The eigenfrequencies and eigenmodes of the footbridge in its initial condition were assessed using the FEM, which was developed based on geometric and material characteristics acquired using the blueprints and bending or flexing experiment outcomes conducted in 1997 [144]. The study reported that the revised FEM precisely reflects the responses of the footbridge and is an appropriate reference FEM for extended tracking to assess the all-around responses in operational loading situations [144].
In the study, Catbas et al. [145] explored vision-based technologies for practical, low-cost and effective methods in non-destructive structural evaluation. A framework for image-based structural identification was proposed to determine structural parameters using a fully non-contact and non-target method based on vision-related techniques [145]. The overview of this non-target deflection acquisition technique is shown in Figure 22. The camera system also identified traffic over the bridge and classification of the vehicle weight estimation [145]. The input effects, including vehicle loadings, vehicle locations, and output displacements of the bridge structures, can be used to determine a series of displacement unit influence surfaces [145]. Quantitative analysis of the unit influence surfaces has confirmed its consistency and reliability, making it suitable for applications in damage detection and localization [145].
In the study by Dong et al. [146], a fully non-contact structural identification system was proposed, primarily aimed at identifying a footbridge’s unit influence line (UIL) in functional traffic conditions. The study solely used cameras and machine vision methods to obtain the placement of traffic as input and deflections as output [146]. The UILs were identified in laboratory experiments and on a footbridge [146]. The bridge user’s loading was predicted using the derived UIL, and the estimated bridge users’ weights were within acceptable ranges [146].
Dong and Catbas [147] proposed a visual structure characterization system for footbridges that integrates visual deflection acquisition with the loading localization of image-based vehicles. The study extracted the UIL as a metric of structural characterization and validated the presented techniques on a pedestrian bridge [147]. The experiment extracted the filtered UIL using fast Fourier transform, which shows the structural behavior of the footbridges and the load capacity was identified [147]. The study reported that the proposed framework is promising and can complement conventional SHM systems.
Zhao et al. [148] used MIDAS Civil software to specify an FEM based on the eigenfrequency acquired from the vibration experiments of a live footbridge together with ML prediction of the structural material parameters. Figure 23 illustrates the two structures of the neural networks. The experiment outcomes indicate that the footbridge model founded on MIDAS Civil software exhibited precision, but did not satisfy the essential requirements until it was updated using the backpropagation technique, which produced results closely aligned with the sensor measurements [148].
Seon Park et al. [149] proposed an extended stress forecast approach for concrete buildings using meteorological information and stress measurements with a CNN. The study used meteorological information as the input and stress measurements from the fiber-optic sensors as the output layer of the CNN [149]. The study reported that it establishes a useful approach for forecasting stress in concrete buildings using the presented method [149].
3.4.21. System Identification
Ali et al. [8] conducted system identification (SI) on a pedestrian bridge by capturing the vibration data of the footbridge with six economical triaxial micro-electromechanical systems (MEMS) accelerometers. Ali et al. [8] used an output-only method of modal analysis, the FDD algorithm, for the derivation of modal characteristics, eigenfrequencies and modes. The study identified three mode shapes and frequencies through system identification and compared them with the bridge’s FEM created with Abaqus software [8]. The study observed a tight association between the FEM and system identification outcomes, with a frequency difference of approximately 10% and a modal assurance criterion (MAC) exceeding 0.80 for experimental and analytical mode shapes [8]. This result demonstrates a close match despite the limited number of accelerometers and the simplifications and idealizations in the FEM [8].
In the study, Duarte and Ortiz [150] performed a viability investigation using economical data collection equipment, smartphones, and Raspberry Pi for system identification using the Bayesian Spectral Density method to determine the model characteristics and the associated unpredictability regarding probability distributions. The approach was verified on a steel pedestrian bridge with a traditional approach of highly sensitive piezoelectric accelerometers for comparison [150]. The results show that all three devices consistently recognized the eigenfrequency, with minimal discrepancies lower than 0.01% mean and variation, thus confirming that economical data collection equipment can accurately determine the structure’s eigenfrequency, yielding results comparable to traditional acquisition systems [150].
Efficiently addressing the multi-parameter recognition challenge for live structures while precisely quantifying ambiguities for determining model characteristics and revising FEM remains a significant challenge [151]. An innovative Bayesian approach utilizing the Markov Chain Monte Carlo method was presented by Liu et al. [151] to tackle these challenges. The proposed framework’s effectiveness was demonstrated through multi-configuration atmospheric vibration experiments on a high-traffic pedestrian bridge, with accelerations measured at specific points [151]. The study reported successfully identifying the prevailing modal behavior and modal characteristics of the pedestrian bridge concurrently with the related ambiguities using the presented method [151].
Most bridge system identification methods currently use output-only approaches, assuming dynamic loads are white noise due to the challenges of measuring them, leading to considerable flaws [152]. Lim and Yoon [152] sought to create a system characterization technique for footbridges that utilized AI and visual recognition methods to estimate the position, extent of the human loading and the vibration reactions of the bridge [152]. The validation outcomes showed that the presented technique successfully evaluated the vibration characteristics of the footbridge with anticipation of serving as an efficient and effective tool for SHM [152].
In the research work, Fábry et al. [153] concentrated on characterizing footbridges by non-invasive dynamic assessment. Several numerical models were developed to facilitate comparison, examining how the rigidity of the parapet affects the overall structure’s rigidity [153]. Outcomes from the experiment revealed that the FBG FS6500 optoelectronic sensor and the PCB393B31 piezo acceleration sensor produced a fairly matching result, thus forming a foundation for the characterization of the footbridge but it is not yet possible to verify structural damage [153].
Donnelly et al. [154] sought to comprehensively investigate the vibration characteristics of the Ha’penny pedestrian bridge by using the associated model in SHM and damage detection. The study initially developed a preliminary finite element model using MATLAB, followed by the use of the modes derived from this analysis to support an OMA conducted with strain gauges attached to the structure [154]. The study reported that after updating and calibrating the FEM, an accurate representation of the bridge is within acceptable limits [154].
3.4.22. Temperature Assessment
Manconi and Moonen [155] presented a numerical system for forecasting temperature in flax-based laminate sandwich pedestrian bridges. 82 FBG devices and eight thermocouples were instrumented to enable real-time evaluation. The designed system is capable of accurately representing the sophisticated thermal spread of the pedestrian bridge with the result of ±1 ÷ 3 °C tolerance range in diverse climatic conditions [155].
Incremental thermal sensing deviations are challenging to notice and will lead to flaws in the temperature correction in stress sensing, potentially being mistaken for time-dependent structural behavior [156]. In their study, Pereira and Glisic [156] employed a probability-based neural network for nonlinear temperature forecasting from data-based models to reduce seasonal systematic deviation. An innovative shift recognition technique established on parameter progression from a three-category multinomial distribution was presented along with a drift characterization approach [156]. The proposed approach was applied to several sensing devices implanted at different places on the deck of the subject pedestrian bridge, showcasing its relevancy in practical settings across different scenarios [156].
3.4.23. Vibration Analysis/Modeling
In this research, Omidalizarandi et al. [157] suggested a reliable and automated dynamic assessment method, a time-based modal analysis recognition technique. This method is innovative in its ability to automatically and reliably identify initial eigenfrequencies, even those that are closely spaced, while reliably and precisely predicting the characterizations of a footbridge with the use of a small number of budget MEMS acceleration sensing device [157]. The dynamic assessment technique demonstrated that magnitudes could be evaluated with precision within a fraction of a millimeter, the precision of frequencies finer than 0.1 Hz, and damping parameters with accuracies finer than 0.1% for modal damping and 0.2% for overall damping [157].
In the study, Casciati et al. [158] presented and analyzed data collected from a pedestrian timber bridge under various “moving” load configurations using a time-frequency decomposition technique to model human-induced vibrations. The motion rates of pedestrians movement on the footbridge were recorded using accelerometers and analyzed to identify the footbridge’s characteristics, followed by the implementation of the numerical model of the pedestrian’s loads [158]. The study reported that the outcomes acquired from the presented model strongly correspond with the field measurements [158].
Kadota et al. [159] conducted a vibration modeling study utilizing 3D acceleration measurements and high-resolution FEM analysis of the joining part of the stairs and main girder of a pedestrian overpass, which exhibited actual damage. The analytical results of this study did not align perfectly with the measured results across multiple natural modes, indicating the need for a thorough review of the analytical framework, including member and material characteristics [159].
Ngan et al. [160] designed an FEM for the assessment of the vibration characteristics of a new pultruded GFRP floor system, which, due to its lightweight and low damping characteristics, is prone to excessive vibrations caused by human dynamic loading. The edge constraints of the end support and beams of the footbridge serve as a solid validation reference for a comparable unidirectional spanning GFRP flooring design [160]. While the verification was conducted on a small collection of results, the investigation highlights the possibility of FE modeling to accurately estimate the vibration properties of the subjected floor design [160].
3.4.24. Vision-Based Modal Testing
In their study, Wang et al. [161] introduced an innovative visual modal assessment framework designed to address challenges such as low measurement accuracy, poor robustness under varying lighting conditions, and the need for artificial target installation. Wang et al. [161] used a multi-configuration approach with 3 time-aligned video capture devices to capture visual images of structure vibrations and a reliable target-tracking method to derive accurate deflection from the captured visual images with substantial noises. This non-contact system is capable of capturing the bends and torsion characteristics of a real bridge under unregulated environments [161]. The study validated the presented technique with an actual-size lab bridge and a live suspended pedestrian bridge, achieving satisfactory modal analysis results consistent with accelerometer measurements, thereby establishing the system’s effectiveness in OMA [161].
3.5. Damage Detection
Damage detection is an essential process in SHM, involving identifying and locating damage by monitoring and analyzing changes in a system’s current state compared to its reference state, which is characterized by specific nominal properties that remain consistent in all healthy conditions [162]. The damage-detection category represents 15.8% of the publications reviewed and consists of four subsections [162].
3.5.1. Algorithm Development
In their study, Jimenez-Alonso et al. [163] presented a novel approach established on complex measurements, referred to as persistent entropy, to address the defect identification challenges. The algorithm’s effectiveness was confirmed through the theorem on the stability of persistent entropy [163]. In the study, Jimenez-Alonso et al. [163] designed, built, and later damaged a laboratory bridge. The outcomes demonstrate the method is responsive and performs well for defect identification [163].
In the study, Yin and Zhu [164] proposed a customized formula for effectively creating the configuration of a Bayesian deep-learning model, optimizing the count of internal neurons and the activation functions for damage detection. Yin and Zhu [164] validated the performance of the presented process by employing stochastic FEM refinement of a footbridge using measured data and the experiment configuration. The results showed that the presented formula optimized the Bayesian deep-learning model architecture and selected the appropriate count of internal neurons and suitable transfer function for both output and hidden layers [164].
Ngoc-Nguyen et al. [165] proposed an innovative technique to address the limitations of the Bee Algorithm by incorporating it with the genetic algorithm for damage detection. The research evaluated an in-service suspended pedestrian bridge for the validation and verification of the significance of the suggested technique [165]. A reference footbridge FEM was created with measurements acquired using the dynamics and model refinement, then employed to create various imaginary failure conditions [165]. The results demonstrated the presented algorithm’s effectiveness in detecting damage across various scenarios [165].
You et al. [166] suggested an Interactive Damage Index Method to enhance the precision of fault identification after reviewing the method. You et al. [166] then numerically and experimentally compares these two damage-detection techniques with data collected from a beam with end supports and a footbridge. The study obtained the vibration characteristics with WSN and the Time Domain Decomposition, while the vibration properties for the quantitative evaluation were obtained from the dynamical assessment using ABAQUS software [166]. You et al. [166] reported that the proposed Interactive Damage Index Method demonstrated greater accuracy and reduced arbitrariness compared to traditional damage-detection algorithms when applied to both structures.
Ellenberg et al. [167] presented a side-by-side evaluation of the sensing abilities of an off-the-shelf UAV, an imaging sensing device TRITOP, and the X-Box Kinect. The proof reveals that UAV-mounted RGB cameras can identify fractures of similar size to those identified during visual assessments, even from varying distances [167]. The study analyzed challenges and errors in marker-based measurements, concluding their feasibility for field conditions, and successfully demonstrated the concept of detecting critical tags for algorithm robustness [167].
Ghasr et al. [168] presented preliminary findings on using microwave synthetic aperture radar (SAR) imaging techniques and a custom-developed SAR imaging algorithm to detect layer separation and the impact of reinforcement rusting on a footbridge with a section visually showing the damage effects. The investigation demonstrated the capability to detect delamination and deterioration caused by moisture infiltration into the concrete of a real pedestrian bridge, highlighting the significant potential for further research [168].
Establishing a dense sensor network with commonly used point sensors or one-dimensional distributed sensors is costly and often impractical [169]. In the research, Kumar et al. [169] derived analytical models and algorithms for fault identification, localization and evaluation built upon a sheet sensor comprising a high-density collection of resistance strain sensor modules. Kumar et al. [169] performed laboratory experiments with simulated fault for validation and used the fault estimation procedure to approximate the gap in a contraction gap on the bridge. The research concluded that the 2D sheet sensor is effective for immediate crack identification, localization, and evaluation using the proposed procedure [169].
Abadía et al. [170] proposed an explainable AI method for SHM to improve practitioners’ confidence in AI techniques. The approach leverages unsupervised learning algorithms to detect outliers that indicate faults in collected measurements and explain the outlier detection process to the user to ensure openness in the decision process [170]. The presented approach as illustrated in Figure 24, validated through pedestrian bridge simulations, effectively differentiates between structural faults and unpredictable changes in the structure’s characteristics [170].
In the study, Kostic and Gül [171] used a time-series evaluation with a sensor clustering technique combined with ANN to detect faults under thermal deviations. Kostic and Gül [171] selected ANNs to account for temperature effects on damage detection, as fault properties derived from the algorithm, while effective, yielded false positives under temperature variations. From the outcomes, the presented technique is useful in determining the presence, place, and comparative intensity of faults with output-based dynamic and thermal measurements, despite being with the existence of thermal deviations [171].
Metaheuristic algorithms that can tackle challenging optimization problems have received significant interest, particularly in damage identification in the field of SHM [172]. In the study, Long et al. [172] proposed creative nature-inspired hunting algorithms that mimic the behaviors of wild animals to determine the places and severity of damage in a suspended pedestrian bridge. The results show that Every algorithm considered in the study accurately detected the damage in the examined structure [172].
In the study, Ngoc-Nguyen et al. [173] presented a novel optimization algorithm, the Slime Mould Algorithm, to address the deterioration identification challenges in a suspended pedestrian bridge. The functionality of the suggested technique was validated using the cuckoo search and genetic algorithm to assess its performance [173]. The outcomes demonstrate that the presented Slime Mould Algorithm surpassed both cuckoo search and genetic algorithm for fault assessment and localization [173].
He et al. [174] proposed two feasible algorithms to optimize the count of neurons of a multiple hidden-layer Bayesian neural network. The algorithms verified with the finite element (FE) model on both a planar truss model and a real-life pedestrian bridge showed one of the algorithms is more computationally efficient with potential for practical application.
3.5.2. Artificial Intelligence
Kostic and Gül [175] proposed a framework that emphasizes damage detection under varying temperature effects. This research employs a time-series clustering assessment approach integrated with ANN to address the influence of thermal deviations on damage detection [175]. The ANNs employed the back-propagation method and the Levenberg–Marquardt algorithm [175]. The study applied the proposed methodologies to an FEM of a pedestrian bridge, demonstrating its effectiveness in successfully determining the presence, spot, and scope of faults under different load cases and varying noise levels [175].
Existing image-based deflection acquisition techniques face restrictions, including the requirement for manual target placement, parameter adjustments, and substantial user intervention to achieve the desired outcomes [176]. In the study, Dong et al. [176] presented an innovative deflection acquisition technique for structures using AI-based whole-field optical flow methods. Dong et al. [176] validated the proposed method’s performance through investigations in the lab on a viewing stand. Statistical assessment of the relative outcomes indicates that the presented approach outperforms the conventional method in accuracy and produces reliable outcomes aligned with deflection data from the sensor [176].
Wootton et al. [177] utilized Echo State Networks (ESNs) to analyze multidimensional, longitudinal time-series data collected from a sensor array for SHM and structural assessments. Wootton et al. [177] used a magnetic flux leakage (MFL) technique to collect an extensive database of MFL signals from a reinforced concrete testbed with artificially induced cracks and corrosion. The study trained two ESNs: one to identify characteristic defect signals in the MFL data and the other to detect the noise in the end regions after the MFL energization process [177]. The study demonstrated that ESNs were capable of processing both spatial and temporal data in detecting structural damage [177].
Wootton et al. [178] applied ESN to chronological measurements collected from a sensor cluster for SHM on an experimental pedestrian bridge, that underwent several possibly harmful actions throughout 3 years. The chronological measurements were collected roughly in 5-minute intervals by 10 thermal sensing devices and 8 inclination sensing devices installed on the pedestrian bridge [178]. The investigation reported that the damaged regions could be identified by comparing the signal errors with the location of each tilt sensor [178]. In a subsequent study, Wootton et al. [179] found that ESNs are effectively applied as classification tools to identify and describe the various kinds of interventions that had occurred [179].
Conventional AI-based damage-detection systems process data on a server [180], whereas Lee et al. [180] proposes a lightweight real-time AI diagnostic system that operates directly on sensor nodes. Tests conducted on a pedestrian bridge demonstrated the effectiveness of the on-device AI in classifying different scenarios.
Li et al. [181] presented a fault identification and classification technique for smartphone data using a 2D CNN. Li et al. [181] reported that the technique with the use of a computational model of the scooter’s actions on a live pedestrian bridge indicates precision in fault identification.
To facilitate rapid estimation and health monitoring of bridge comfort levels, Zhao et al. [182] integrated finite element simulation results to develop a data-driven library and introduced three distinct ML algorithms. The results confirmed the model’s effectiveness in quickly assessing bridge comfort levels, providing valuable insights for bridge health monitoring.
3.5.3. Method Development
Ahmadi and Anvari [183] presented a novel approach to detect faults in pedestrian truss bridges established on a cone-like kernel function and a novel fault metric. Ahmadi and Anvari [183] designed the proposed algorithm that eliminates the requirement to construct a mathematical model of the subject. The results show that the suggested procedure and fault metric effectively recognized the faults and accurately pinpointed their position [183].
In the study, Zhang and Sun [184] sought to merge the advantages of data-based and physics-driven SHM techniques through physics-guided ML to enhance fault recognition reliability. The study reported that the experiments on a 3-level frame structure demonstrated the usefulness of physics-guided ML in identifying damages, whether using data-based methods or FEM refinement techniques [184].
Sigurdardottir and Glisic [185] adopted a holistic approach to developing new fiber-optic techniques for SHM established on stress measurements to gain the fourth level in SHM. The outcomes encompass initial phase and extended stress progression, fault identification and characterization, assessment of joint rigidity reduction, model assessment, and assessment of complicated cross-section effectiveness [185]. This investigation demonstrated that the reliability of the techniques can reach Level IV inspection [185].
Yaghoubzadehfard et al. [186]’s research focused on a non-contact approach using a radar interferometry device, the IBIS-FS, which is renowned for its mobility and non-invasive function. This study used IBIS-FS to capture the eigenfrequencies and vibrations modes of a footbridge, with the results aligning well with those from FEM, thereby establishing the trustworthiness of IBIS-FS in identifying modal characteristics [186]. Figure 25 illustrates the setup of the IBIS-FS approach. The research highlighted the possibility of integrating advanced AI methods using measurements acquired by IBIS-FS [186].
Variations in measured natural frequencies can occur due to alterations in the environment, structure and soil interaction, or earthquakes, complicating the determination of whether a localized variation in a structure is a result of damage or other influencing elements [187]. In the study, Liao and Petryna [187] proposed the wave transmission assessment technique and the transmittance correlation to overcome the issue, therefore enhancing the present fault identification technique. The study reported that the structural changes matched the proposed damage indicators [187].
Pepi et al. [188] presented a robust fault recognition framework established with Bayesian modeling of ambiguous variables in the FEM used for vibration assessment. Pepi et al. [188] evaluated the significance with virtual response samples contaminated with measurement noise. The framework explicitly addresses measurement errors in the Bayesian updating process to solve damage identification and localization problems [188].
3.5.4. Model Development
Modal strains, derived from dynamic strain measurements, are modal characteristics that have proven to increase responsiveness to localized damage and are negligibly affected by thermal conditions compared to eigenfrequencies [189]. In the study, Anastasopoulos et al. [189] dynamically excite the footbridge using hammer impacts on the footbridge Fiber Reinforced Polymer (FRP) deck made out of one piece by vacuum infusion to perform experimental modal analyses. Anastasopoulos et al. [189] used embedded FBG stress-sensing devices to monitor the footbridge’s strains under dynamic loading. The study used the identified modes to update an FEM of the footbridge constructed using ANSYS software and simulated possible damage scenarios that can occur on FRP structures in the FEM, and investigated the impact of these on eigenfrequencies and modal strains of the footbridge [189].
3.6. Reporting
Reporting is a workflow of collecting, organizing, and presenting information to communicate specific insights, findings, or updates to a targeted audience, which involves structuring data into formats such as text, tables, charts, and graphs to make the information clear and actionable. Reporting can serve various purposes, including monitoring performance, supporting decision making, and ensuring accountability across diverse fields. The reporting category represents 4.7% of the included literature.
3.6.1. Deflections Documenting
In the study, Stiros and Moschas [190] presented what is likely the earliest instance of continuous deflection assessments documenting the deterioration of a bridge’s structural health over an extended period and with significant intensity. Measurements of sideway displacements of a wooden footbridge illustrated in Figure 26, induced by pedestrian traffic, collected annually with a robotic surveying instrument in the span of six years, revealing a decrease of 1.6 Hz eigenfrequencies through 2007 and 2008, followed by an incremental decline of about 8% from 2008 to 2012 [190]. The frequency decreases corresponding to evidence of deterioration with a decrease in eigenfrequencies in the vertical orientation likely triggered by a significant earthquake, particularly an unusual icing event [190].
3.6.2. System Performance
In their publication, Xi et al. [191] reported and documented the system performance from the assessment of a metropolitan pedestrian bridge. Xi et al. [191] analyzed and assessed the operation performance of the footbridge using strain and deflection measurements collected by FBG sensing devices deployed at crucial positions. Xi et al. [191] used a linear modeling technique to isolate the thermal impact from the measurements. The pedestrian bridge’s vertical deviations were compared with the vibration assessments acquired by an image-based device from a distance [191].
Xia et al. [192] presented a brief review of the SHM implemented for a butterfly-arch pedestrian bridge. Xia et al. [192] reported that the system comprises a Brillouin optical sensing system, FBG, GNSS, and high-resolution cameras. The manufactured fire incident used to showcase the system’s functionality and significance offered a unique chance for fault detection in an operational footbridge [192]. The SHM tracks the bridge’s functionality on-the-fly and offers cloud-enabled software for investigation and education, as well as a touchscreen visualization platform for public outreach [192].
3.6.3. System Valuation
Nepomuceno et al. [193] proposed to test Vardanega et al. [194]’s value assessment process that offers a systematic approach for the valuation of an SHM strategy before its implementation. This process produces a straightforward measure that evaluates the likelihood of an SHM strategy delivering value to stakeholders [193]. An FRP pedestrian bridge was utilized to validate this methodology, comparing an accelerometer setup and a strain gauge setup [193]. The methodology effectively facilitated discussions among core participants at the initial stages [193].
3.6.4. Test Results
In the paper, Buffarini et al. [195] presented the outcomes of dynamic assessment of a footbridge. The footbridge suffers damage and decay caused by environmental influences [195]. The investigation focuses on the tallest columns that exhibited irregular behavior during an initial exploratory work [195]. The outcomes indicated that the bridge’s behavior was qualitatively consistent with the expected outcomes [195].
In two different studies, Kulpa and Siwowski [196] and Kulpa and Siwowski [197] presented test results in SHM of footbridges made with fiberglass-reinforced plastic (FRP) established on DFOS technology. The installed DFOS technology demonstrated a strong correlation with conventional data collected from the load testing. The subject will be continuously observed and the results collected can be compared with subsequent results to determine the behavior over time [196,197].
Marasco et al. [198] conducted thorough research on the applicability of unsupervised domain adaptation for assessing the structure’s stability of comparable structural designs. Test results and insights from structures are transferred across domains to assist authorities in improving decision making [198]. Marasco et al. [198] anticipates that the method will enhance classification accuracy in the target domain by reducing distribution mismatches.
4. Discussion, Challenges, and Directions for Future Research
4.1. Synopsis of the Scientometric Study
Research on SHM for footbridges began to emerge nearly two decades ago. Over the past decade, there has been a noticeable upward trend in publications, increasing from 6 papers per year in 2014 to 22 papers per year in 2023, highlighting the growing interest in SHM for footbridges. Figure 2 illustrates the trends of SHM in footbridges. The publications span book sections, conference papers, and journal articles, with conference papers and journal articles having nearly equal distribution at 44.4% and 53.8%, respectively, as shown in Table 1. While journal articles constitute a significant portion of publications of the SHM for footbridges, the International Conference on Structural Health Monitoring of Intelligent Infrastructure stands out as the leading source publications, as highlighted in Table 3. Table 2 and Table 3 provide a valuable reference for emerging researchers in SHM for footbridges to select suitable publishers for their work.
This review explored the coexistence of keywords to determine the core articles of the published papers. The keyword mapping results, presented in Section 2.2, highlight the number of times the keyword appears, showcasing the most commonly used terms in this field of study. After excluding generic keywords such as SHM, footbridges, monitoring, etc., the analysis of keywords listed in Table 4 revealed that the highly researched keywords identified from the publications are modal analysis, damage detection, fiber optic sensors, structural dynamics and accelerometers. At the same time, keywords like computer vision, modal identification, wireless sensor networks, neural networks, and operational modal analysis require further exploration.
The co-authorship analysis in this review examined the connections between prominent researchers, institutions, and the countries where the research originated. The co-author relationship visualization presented in Section 2.3 and Section 2.4 highlighted worldwide interactions between respected authorities and leading organizations. Based on Figure 4, Glisic, B., is the central and most influential author, connecting all major groups in the network. The Red Cluster, consisting of Kumar, V., Sturm, J.C., Verma, N., Aygun, L.E., etc., appears to be a core team from Princeton University working closely connected with each other and Glisic. The Green Cluster formed by Domaneschi, M., Cimellaro, G.P., etc., from Politecnico di Torino, Italy, was collaborating with Glisic’s team. Abdel-Jaber, H. and Morris, I., who form the Purple Cluster, are collaborating with Glisic, who is also affiliated with Princeton University. The Blue Cluster, which includes Park H.S., Oh B.K., Lee D.E., and Hong T. from Yonsei University in Korea, reflects an international collaboration with Glisic. Napolitano, R. and Blyth, A., also affiliated with Princeton University, make up the Cyan Cluster, a smaller group collaborating with Glisic. Finally, Pereira, M., also from Princeton University, works closely alongside Glisic. In summary, Glisic collaborates with four teams from Princeton University, as well as with teams from Politecnico di Torino in Italy and Yonsei University in Korea. The co-authorship analysis, along with the data presented in Table 5, indicates that the Streicker Bridge at Princeton University has made the most significant contribution to research publications in this field. The USA, China, and Italy are the top three contributing countries, accounting for approximately 57.9% of the total publications. However, the USA has established strong research connections with Italy, the United Kingdom, Spain, South Korea, and Ireland. On the other hand, countries with weaker research ties, as shown in Figure 5, include Australia, Canada, and Vietnam.
Although the VOSviewer software can extract keywords, including sensor types mentioned in each publication, it lacks the ability to scan the entire text of the publication to identify the specific sensors used. As a result, this review manually read and analyzed each publication, extracting the types of sensors used and compiling the information into Table 7. The results indicate that accelerometers, cameras, and fiber optic sensors are the most commonly used sensors.
4.2. Challenges
4.2.1. Cost of the SHM System
Approximately 17.5%, or 30 papers, have highlighted that the cost of instrumentation, data acquisition systems, installation, cabling, data storage, data analysis, maintenance, operational expenses, energy and labor represent significant obstacles to the deployment of SHM in footbridges, preventing government agencies and bridge owners from adopting such systems. These research either propose cost-reduction methods or focus on developing low-cost systems to address the financial challenges associated with SHM deployment [8,18,24,26,31,32,33,34,37,40,42,47,53,59,73,83,84,86,94,107,116,129,130,132,141,145,146,150,157,166,169,173]. A key solution emphasized in these papers is the development of a non-contact vision-based system, as highlighted by Xu et al. [84], Catbas et al. [145], Omidalizarandi et al. [83], Dong et al. [146], Fradelos et al. [47], Komarizadehasl et al. [132], and Bai et al. [94]. This review will delve deeper into contactless visual-based systems in Section 4.3.
4.2.2. Data Transmission
Section 4.2.4 described challenges identified in data transmission for wireless systems. However, even a wired system encounters the same challenges [143,149,184]. As a result, accurate and error-free data transmission remains a key challenge for SHM of footbridges, regardless of whether wired or wireless systems are used. Despite the research discussed in Section 4.2.4 and Section 4.2.2 focusing on data correction algorithms [78,79,80,81], this review did not identify any studies explicitly addressing the root cause of the issue, the data transmission medium.
We conducted a Scopus database search using the keywords (“data transmission” AND integrity) for the period 2014 to 2024 to gain an overview of research in this area. The search returned 1,481 publications; however, when refined with the addition of “SHM”, only 12 relevant papers were identified, representing just 0.8% of the total, indicating limited focus on data transmission integrity within SHM research.
Data transmission integrity is crucial in SHM for several reasons. Many footbridges are located at remote or hazardous locations, for example, mountain footbridges, river footbridges, and footbridges over heavy traffic. Lost or corrupted data often necessitates manual retrieval or system resets, which can be costly, time-consuming, and potentially hazardous. The study by Satme et al. [27] presents an innovative method for the sensor deployment over these remote locations; however, it still requires a human operator to operate the UAV and manually retrieve the data. For critical bridges, such as multi-use bridges shared by vehicles, rail, and pedestrians like the Sydney Harbour Bridge, or rail-pedestrian shared bridges like the Hutt River Bridge in New Zealand, real-time SHM is essential for timely safety alerts. When data integrity is compromised, faults may go undetected, increasing the risk of delayed interventions and potential safety hazards. In Australia, floods over bridges are common. After floods, SHM sensors could potentially provide data on damage levels. Corrupted or missing data during transmission could result in incorrect assessment of structural safety after floods, putting responders and the public at risk. In long-term SHM, even minor corrupted data spanning over years or decades can accumulate, degrading the reliability of trend analysis used for damage prediction.
Data integrity is not solely a computing concern; in SHM, it directly impacts public safety, maintenance costs, and infrastructure resilience. Therefore, more robust communication protocols, redundancy, and standardized error-checking methods are needed for future SHM designs and warrant future research focus.
4.2.3. Integration of Multi-Source Data (Data Fusion)
SHM systems are often deployed with a combination of sensing devices, like accelerometers, strain transducers, cameras, fiber optics, and temperature sensors. According to the data in Table 8, around 26.9%, or 46 papers, have employed multiple sensors in their studies. However, integrating data from multiple sensor types into coherent, actionable insights remains challenging, particularly when dealing with different sampling rates and data formats [67]. Researchers addressing these challenges include Ozer and Feng [67] (smartphone-based sparse SHM data), Ponsi et al. [111], Moschas and Stiros [38], and Casciati et al. [107] (GNSS and accelerometer fusion), Mustapha et al. [105] (FBG with accelerometers), Quqa et al. [64] (multi-smartphone sensors), and Ma et al. [86] (accelerometer and FMCW integration). Only a small proportion, about 10.9% (5 studies), of the 46 publications that employ multiple sensors focus on fusing data from these different sensor sources, highlighting a significant research gap.
Future research should not only, focus on the advancement of data-fusion techniques such as Kalman filtering [107,111], machine-learning-based sensor fusion [105], and finite impulse response filter [86], but also aim to standardize protocols for time alignment and synchronization [52,79,123,199] and harmonization of noise characteristics across sensor types [200]. These research areas could enable advanced SHM systems capable of predictive maintenance and automated damage detection.
4.2.4. WSN Accuracy, Synchronization, Data Loss, and Energy
Citing the high cost of a wired SHM installation, researchers have also proposed low-cost wireless sensor networks as alternatives. Sabato [33], Navabian and Beskhyroun [32], Komarizadehasl et al. [31], Komarizadehasl et al. [132] and Bai et al. [94] are all working on low-cost wireless solutions. However, whether commercially manufactured or custom-built in laboratories, wireless systems, though regarded as a cost-effective alternative, pose additional challenges due to their distributed and wireless characteristics. These challenges include maintaining data accuracy during transmission [27], achieving synchronization for data processing [52,79], and data loss [78]. Another significant challenge for a wireless sensor network (WSN) is its energy source. Table 9 identifies 29 publications utilizing wireless systems as the data transmission medium. However, only one study, Jiménez-Alonso et al. [19], explored the feasibility of powering WSNs with energy harvested from pedestrian-induced vibrations, revealing a significant research gap that underscores the need for further investigation and advancement in this area.
4.3. Future Directions
This analysis and review of the SHM of footbridges publications from the past decade has uncovered several emerging and increasingly evident trends. During this time, researchers have concentrated on addressing gaps in the field of SHM for footbridges by introducing innovative technologies, algorithms, methods, and sensors.
4.3.1. Artificial Intelligence
The COVID-19 pandemic accelerated the deployment of artificial intelligence (AI) across various domains in early 2020 [201]. Correspondingly, research on SHM of footbridges utilizing AI has notably increased since 2020, as shown in Figure 27. In Section 3.5.2, this review discussed research that focused mainly on AI for damage detection. However, according to other sections of this review, other researchers have also utilized AI as an analysis tool to assist in interpreting experimental results [51,58,76,104,105,115,141,143,148,149,156,164,170,171,174,184].
Advances in explainable AI could improve transparency, trust and interpretable decision making for stakeholders [202,203]. Furthermore, integrating AI with emerging technologies such as TinyML using lightweight CNN models with low-power microcontrollers [204] and edge computing enables efficient real-time processing with reduced latency and lower energy consumption [205]. These approaches potentially align with energy harvesting techniques by Jiménez-Alonso et al. [19] and may help address energy source challenges discussed in Section 4.2.4.
Recent developments in AI, including digital twins and deep learning using generative adversarial networks (GAN), particularly data fusion [206] and data augmentation [207] in road bridges, have highlighted their potential for footbridge applications. As these technologies continue to advance, particularly in areas such as ANN [171], CNN [143,204], and SVM, they have the potential to shift the SHM field from run-to-failure maintenance toward preventive maintenance strategies, greatly improving the safety and cost-effectiveness of footbridge maintenance.
To realize these advancements, future research should prioritize real-time on-board AI processing methods [208,209,210] to enhance decision-making speed rather than relying on cloud processing. Sensor fusion using ML [105,211,212] should also be explored to address gaps discussed in Section 4.2.3 along with anomaly detection with ML [6,76,156]. Cross-disciplinary collaborations will be essential to accelerate the adoption of these technologies in SHM of footbridges.
4.3.2. Crowd Sensing
Crowdsensing, or crowdsource-based SHM, is an emerging area that is gaining increasing attention [62,63,64,65,66,67]. Crowdsensing systems differ from conventional sensor systems by functioning as a dynamic, mobile, and distributed network of sensor components [63]. This approach commonly utilizes various smartphone sensors, such as accelerometers for capturing vibration data [63,80], GPS for measuring displacement [65,121], and timestamp [105,121], Wi-Fi or cellular networks for data transmission [66,67], and the smartphone’s central processing unit for on-device computing and data analysis [67].
As reported by the publications in Section 3.3.1, crowdsensing techniques share common challenges with WSNs due to similar characteristics such as mobility, energy constraints, inconsistent sampling rates, and data synchronization issues. Despite these challenges, smartphones are unlikely to face barriers in their adoption for SHM of footbridges, as the smartphone industry is anticipated to experience growth to approximately USD 792 billion before 2029 [213]. The rapid evolution of cloud computing [214], such as Amazon Web Services, Google Cloud, and Microsoft Azure, facilitates seamless integration with smartphone applications, further accelerating the development of smartphones as low-cost SHM solutions [215,216,217].
Thus, crowdsource-based SHM, a publicly funded solution involving contributions from citizens, holds significant potential to address cost-related challenges faced by governments and bridge owners, and is expected to receive increasing attention from the research community.
4.3.3. Low-Cost SHM Technologies
As highlighted in Section 4.2.1, cost remains a major barrier for governments and asset owners seeking to implement SHM on footbridges, particularly for ageing or underfunded infrastructure. In response, researchers are increasingly focusing on low-cost SHM solutions, leveraging recycled e-waste [71], widely available MEMS accelerometers [8,32,33,59], off-the-shelf cameras [84,145], and smartphones [150]. These technologies significantly reduce deployment costs and offer acceptable performance (1.28% maximum error when compared with a commercial HI-INC solution) [31] for SHM applications.
The recent boom in hobbyist open-source hardware has encouraged the wide adoption of low-cost embedded boards such as Arduino [23,31,59,123,132,204,215,218] and Raspberry Pi [23,31,71,132,150] for SHM applications on footbridges. Open-source software is experiencing tremendous growth across various fields, lowering barriers to access for high-quality, freely available tools. Platforms such as Python [83], OpenCV (computer vision library) [146], FreeRTOS (real-time operating system) [56], Ncorr (2D digital image correlation MATLAB program) [139], and pvlib-python (Python photovoltaic library) [155] have made advanced functionalities freely accessible to researchers in the SHM domain.
As the open-source ecosystem continues to expand, SHM research stands to benefit substantially from the growing availability and interoperability of these tools. As open-source hardware and software become more powerful and mature, low-cost SHM solutions are expected to become a major focus in future research and scale toward broader adoption, especially when integrating with AI and cloud computing for real-time analysis, predictive maintenance, and large-scale deployment.
4.3.4. Standardizing Reporting Protocols for SHM of Footbridges
During the review process, we found that key information such as sensing accuracy, deployment difficulty, applicable scenarios, and structural characteristics of footbridges were often omitted or inconsistently reported, limiting the interpretability and applicability of the findings. To advance the field of SHM, we advocate for the adoption of improved reporting standards that include detailed descriptions of the monitored structures. Key structural attributes, such as material type, structural system, span length, age, usage conditions, and climate exposure conditions, are essential for enabling meaningful comparisons across studies and for replicating or extending prior research. Standardized reporting protocols would not only enhance metadata quality for future reviews and scientometric analysis but also support more informed design and deployment of SHM technologies tailored to specific bridge types. Journals and researchers should treat these structural descriptors as essential components of SHM methodology and standard reporting practice. We recommend that journals and researchers treat these structural descriptors as essential components of SHM methodology.
4.3.5. Synthetic Aperture Radar
Synthetic aperture radar (SAR) is a remote-sensing technology that utilizes microwave radar to generate high-resolution images. Although it was mentioned in only one study [168] within this review, SAR, particularly Interferometric SAR (InSAR) and Multi-Temporal InSAR (MT-InSAR), have gained increasing attention in broader SHM applications beyond footbridges, such as road infrastructure, roadway bridges [219], and concrete girder bridges [220]. Figure 28 and Figure 29 illustrate the growing trend of InSAR and MT-InSAR research.
Despite its advantages, SAR technology is not without its limitations; for example, it requires line-of-sight for normal operation [219,220] and a low sampling rate [220]. Studies have shown that while these technologies cannot fully replace conventional SHM systems [221], they hold significant potential as complementary tools in hybrid SHM systems combining SAR and conventional sensors [219,222].
SAR technology offers a cost-effective long-term monitoring solution for early deterioration due to its large-area coverage, non-contact sensing, and suitability for visual inspection in hard-to-access areas [219]. As satellite constellations and UAV-based SAR systems become more accessible [223], their integration into routine SHM workflows is likely to increase.
4.3.6. Vision-Based Technologies
The high cost of conventional SHM systems has motivated researchers to explore alternative technologies, giving rise to several emerging trends. The use of cameras as sensors has emerged as a prominent approach. Research on cameras as non-contact vision-based solutions is driven by two primary factors: low cost [84,94,95,145] and ease of deployment [49,100,102]. Although data from Table 7 shows that cameras and imaging devices are among the most commonly used sensors in the collected publications, vision-based technologies still face challenges, such as low frame rates [50,94], low resolution [46,47,100], camera shake, variable lighting, and limited field of view [95].
Despite these limitations, the strong advantages of low cost and ease of deployment continue to drive interest in vision-based SHM solutions for footbridges. As computer vision technologies, particularly open-source OpenCV [146], and camera hardware continue to improve, vision-based systems are expected to play a larger role in SHM of footbridges.
4.3.7. Wireless Sensor Network
WSNs, introduced as a cost-effective alternative to wired sensor systems, have garnered significant attention since 2018, with more studies noted from 2022 onward [224]. They offer several notable advantages, including minimal cabling costs [116], ease of deployment [32], and scalability [32,67,116]. However, WSNs also present several challenges, such as data loss, energy consumption, and data synchronization, as discussed in Section 4.2.4 and Section 4.2.2. Nevertheless, these issues have not deterred research impact; instead, the cost benefits continue to attract research attention, especially since there are current studies working on addressing these issues, such as Jiménez-Alonso et al. [19] working on energy harvesting, data loss by Fan et al. [225], and wireless data synchronization by V. Shajihan et al. [52] and Dragos et al. [79].
As WSN technologies advance, the integration with AI [226], the rapid growth of smartphones [213], the evolution of cloud computing [214], and the expansion of the open-source ecosystem will contribute to the widespread adoption of WSN systems and improved real-time monitoring and autonomous warning systems. Therefore, WSNs are anticipated to remain a growing area of interest in the field of SHM for footbridges.
5. Conclusions
This study reviewed the SHM of footbridges to extract insights from recent literature. A scientometric analysis was conducted using VOSviewer and custom R scripts on a dataset of 171 selected publications between 2014 and 2024 (from an initial pool of 702). Subsequently, each paper was individually analyzed to examine the current state of research and global trends in SHM for footbridges. The review presents its conclusions in two parts: the scientometric analysis and the state-of-the-art review.
The scientometric analysis of 171 papers published between 2014 and 2024 reveals several key trends in the field of SHM of footbridges. Over this period, the number of publications increased by approximately 433.3%, reflecting growing interest and research activity. Journals accounted for the majority of these publications, comprising 53.8%, which is 9.4% more than those published in conferences. The United States emerged as the leading contributor, producing 28.7% of the total selected papers. Notably, Glisic, B., from Princeton University, in collaboration with internal colleagues and international teams from Politecnico di Torino in Italy and Yonsei University in Korea, formed the largest research cluster, contributing to around 11.7% of the publications. In terms of sensor usage, accelerometers were the most commonly employed, featuring in approximately 50.9% of the studies. Furthermore, about 66.7% of the papers utilized a single sensor, while 26.9% incorporated two or more sensors. Among the research categories, data processing and analysis stood out as the dominant focus, accounting for roughly 50.3% of the total studies.
The state-of-the-art review highlights several key findings. Around 17.5% of the studies identified cost as the primary barrier to the broader adoption of SHM in footbridges. Data corruption remains a significant issue for both wired and wireless systems, hindering accurate and error-free transmission. Although multiple sensors were used in some studies, only about 10.9% applied data-fusion techniques, indicating considerable potential for further exploration in this area. WSNs face persistent challenges, particularly regarding data accuracy, loss, synchronization, and, most notably, energy supply, pointing to a critical research gap. The integration of artificial intelligence in damage detection and analysis has seen a 400% increase, especially during the COVID-19 period, and is likely to draw continued attention. Crowdsource-based SHM solutions using smartphones are gaining momentum, supported by the rapid growth of the smartphone market and advancements in onboard sensors. As cost remains a commonly cited limitation, low-cost SHM solutions using accessible technologies such as MEMS sensors, cameras, and smartphones are expected to attract increasing interest. Furthermore, despite issues like low frame rates, limited resolution, camera shake, and lighting conditions, non-contact vision-based technologies are appealing for their affordability and ease of deployment, making them a promising direction for future research. Overall, these findings suggest a future toward intelligent, decentralized, and low-cost SHM solutions, driven by innovations in AI, sensing technologies, and mobile platforms.
Author Contributions
Conceptualisation, J.L. and M.R.; methodology, J.L. and M.R.; sofware; J.L.; validation, J.L., M.R., K.L., A.M.N. and E.S.; supervision, M.R., K.L., A.M.N. and E.S.; writing—review and editing, J.L. and M.R. All authors have read and agreed to the published version of the manuscript.
Funding
The research received no external funding.
Conflicts of Interest
The authors declare no conflicts of interest.
Abbreviations
The following abbreviations are used in this manuscript:
| AE-FDD | automated enhanced frequency domain decomposition |
| AI | artificial intelligence |
| ANN | artificial neural network |
| CNN | convolutional neural network |
| CPEH | cantilevered piezoelectric energy harvester |
| DFOS | distributed fiber optic sensing |
| ESN | echo state network |
| FBG | fiber Bragg grating |
| FEA | finite element analysis |
| FEM | finite element method |
| FRP | fiberglass reinforced plastic |
| GFRP | glass fiber-reinforced polymer |
| GNSS | global navigation satellite system |
| GPR | ground penetrating radar |
| IATS | image-assisted total station |
| ICSF | identification of combined sensor faults |
| LARA | low-cost adaptable reliable accelerometer |
| LiDAR | light detection and ranging |
| MAC | modal assurance criterion |
| MEMS | microelectromechanical systems |
| MFL | magnetic flux leakage |
| ML | machine learning |
| NRTK | network-based real-time kinematic |
| OMA | operational modal analysis |
| PCA | principal component analysis |
| PPEH | piezoelectric energy harvester |
| PVDF | polyvinylidene fluoride |
| PZT | lead zirconate titanate |
| RTK | real-time kinematic |
| SHM | structural health monitoring |
| SOBI | second-order blind identification |
| SSI | stochastic subspace identification |
| SVM | support vector machine |
| TMD | tuned mass dampers |
| TVF-EMD | time-varying filtering empirical mode decomposition |
| UAV | unmanned aerial vehicle |
| UIL | unit influence line |
| VBI | vehicle-based interaction |
| VR | virtual reality |
| VVS | virtual video sensors |
| WSN | wireless sensor networks |
References
- ASCE. Bridges. In ASCE’s 2021 Infrastructure Report Card; ASCE: Reston, VA, USA, 2021. [Google Scholar]
- Rashidi, M.; Samali, B.; Sharafi, P. A new model for bridge management: Part A: Condition assessment and priority ranking of bridges. Aust. J. Civ. Eng. 2016, 14, 35–45. [Google Scholar] [CrossRef]
- Salvatore, W.; Uva, G.; Venanzi, I.; Mazzotti, C.; Morici, M.; Natali, A.; Dall’Asta, A.; Ubertini, F.; Mannella, P.; Lepori, L.; et al. Application of Italian Guidelines for structural-foundational and seismic risk classification of bridges: The Fabre experience on a large bridge inventory. Procedia Struct. Integr. 2024, 62, 1–8. [Google Scholar] [CrossRef]
- Abdoli, F.; Rashidi, M.; Wang, J.; Siddique, R.; Nasir, V. Structural health monitoring of timber bridges—A review. Results Eng. 2024, 24, 103084. [Google Scholar] [CrossRef]
- Sofi, A.; Jane Regita, J.; Rane, B.; Lau, H.H. Structural health monitoring using wireless smart sensor network—An overview. Mech. Syst. Signal Process. 2022, 163, 108113. [Google Scholar] [CrossRef]
- Noori Hoshyar, A.; Rashidi, M.; Yu, Y.; Samali, B. Proposed Machine Learning Techniques for Bridge Structural Health Monitoring: A Laboratory Study. Remote. Sens. 2023, 15, 1984. [Google Scholar] [CrossRef]
- Moutinho, C.; Magalhaes, F.; Caetano, E. Analysis of the vibration levels of a slender footbridge measured by a continuous dynamic monitoring system. In Proceedings of the Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, Hersonissos, Crete Island, Greece, 25–27 May 2015; pp. 2600–2610. [Google Scholar] [CrossRef]
- Ali, A.; Sandhu, T.Y.; Usman, M. Ambient vibration testing of a pedestrian bridge using low-cost accelerometers for shm applications. Smart Cities 2019, 2, 20–30. [Google Scholar] [CrossRef]
- Mousavi, V.; Rashidi, M.; Mohammadi, M.; Samali, B. Evolution of Digital Twin Frameworks in Bridge Management: Review and Future Directions. Remote. Sens. 2024, 16, 1887. [Google Scholar] [CrossRef]
- Haghani, M. What makes an informative and publication-worthy scientometric analysis of literature: A guide for authors, reviewers and editors. Transp. Res. Interdiscip. Perspect. 2023, 22, 100956. [Google Scholar] [CrossRef]
- Scopus. Scopus Now Includes 90 Million + Content Records! Elsevier Scopus Blog: Amsterdam, The Netherlands, 2023. [Google Scholar]
- Corrin, L.; Thompson, K.; Hwang, G.J.; Lodge, J.M. The importance of choosing the right keywords for educational technology publications. Australas. J. Educ. Technol. 2022, 38, 1–8. [Google Scholar] [CrossRef]
- Van Eck, N.J.; Waltman, L. Visualizing Bibliometric Networks. In Measuring Scholarly Impact; Ding, Y., Rousseau, R., Wolfram, D., Eds.; Springer International Publishing: Cham, Switzerland, 2014; pp. 285–320. [Google Scholar] [CrossRef]
- Rashidi, M.; Mohammadi, M.; Sadeghlou Kivi, S.; Abdolvand, M.M.; Truong-Hong, L.; Samali, B. A Decade of Modern Bridge Monitoring Using Terrestrial Laser Scanning: Review and Future Directions. Remote. Sens. 2020, 12, 3796. [Google Scholar] [CrossRef]
- MRA. Structural Health Monitoring Market By Technology (Wired Structural Health Monitoring, & Wireless Structural Health Monitoring), Offering (Sensors, & Hardware), Vertical (Civil Infrastructure, & Bridges), By Geographic Scope & Forecast 2024–2031. 2024. Available online: https://mraccuracyreports.com/marketreports/5/837363/Structural-Health-Monitoring-Market-By-Technology-Wired-Structural-Health-Monitoring-and (accessed on 27 July 2025).
- Napolitano, R.; Blyth, A.; Glisic, B. Virtual environments for structural health monitoring. In Structural Health Monitoring 2017: Real-Time Material State Awareness and Data-Driven Safety Assurance—Proceedings of the 11th International Workshop on Structural Health Monitoring, IWSHM 2017; IWSHM: Stanford University, CA, USA, 2017; pp. 1549–1555. [Google Scholar] [CrossRef]
- Napolitano, R.; Blyth, A.; Glisic, B. Spherical imaging and virtual environments for structural health monitoring. In Proceedings of the International Conference on Structural Health Monitoring of Intelligent Infrastructure, Brisbane, Australia, 5–8 December 2017; Mahini, S., Mahini, S., Chan, T., Eds.; ISHMII: Brisbane, Australia, 2017; pp. 398–404. [Google Scholar]
- Luleci, F.; Li, L.; Chi, J.; Reiners, D.; Cruz-Neira, C.; Catbas, F.N. Structural health monitoring of a foot bridge in virtual reality environment. Procedia Struct. Integr. 2022, 37, 65–72. [Google Scholar] [CrossRef]
- Jiménez-Alonso, J.F.; López, G.C.; Sánchez, F.G.; Sáez, A. Energy harvesting from pedestrian-induced vibrations in footbridges with piezoelectric devices: A feasibility study. In Proceedings of the Bridge Safety, Maintenance, Management, Life-Cycle, Resilience and Sustainability; Casas, J.-R., Frangopol, D.M., Turmo, J., Eds.; CRC Press: Boca Raton, FL, USA, 2023; pp. 704–712. [Google Scholar] [CrossRef]
- Sieńko, R.; Bednarski, Ł.; Howiacki, T. Smart composite rebars based on DFOS technology as nervous system of hybrid footbridge deck: A case study. In Proceedings of the European Workshop on Structural Health Monitoring; Springer: Cham, Switzerland, 2021; Volume 128, pp. 331–341. [Google Scholar] [CrossRef]
- Bednarski, Ł.; Sieńko, R.; Howiacki, T.; Badura, K. Thermal compensation of monolithic distributed fibre optic sensors: From the lab to the field. Measurement 2024, 238, 115280. [Google Scholar] [CrossRef]
- McSweeney, N.; Ghiasi, R.; Malekjafarian, A.; Ozer, E. Extracting bridge modal frequencies using stationary versus drive-by modes of smartphone measurements. Infrastructures 2024, 9, 218. [Google Scholar] [CrossRef]
- Komarizadehasl, S.; Xia, Y.; Komary, M.; Lozano, F. Eigenfrequency analysis of bridges using a smartphone and a novel low-cost accelerometer prototype. Front. Struct. Civ. Eng. 2024, 18, 202–215. [Google Scholar] [CrossRef]
- Fiandaca, D.; Denoël, V.; Pirrotta, A. A Vehicle Bridge Interaction based approach for the monitoring of bridges through an electric mobile platform. J. Phys. 2024, 2647, 182016. [Google Scholar] [CrossRef]
- Li, S.; Xu, C.; Pan, W.; Pan, Y. Sensor deployment optimization for detecting maneuvering targets. In Proceedings of the 2005 7th International Conference on Information Fusion, Philadelphia, PA, USA, 25–28 July 2005; Volume 2, p. 7. [Google Scholar] [CrossRef]
- Gong, Z.; Han, L.; An, Z.; Yang, L.; Ding, S.; Xiang, Y. Empowering smart buildings with self-sensing concrete for structural health monitoring. In Proceedings of the SIGCOMM 2022-Proceedings of the ACM SIGCOMM, Amsterdam, The Netherlands, 22–26 August 2022; pp. 560–573. [Google Scholar] [CrossRef]
- Satme, J.N.; Yount, R.; Smith, J.; Downey, A.R.J. Case study for using open-source UAV-deployable wireless sensor nodes for modal-based monitoring of civil infrastructure. In Proceedings of the 14th International Workshop on Structural Health Monitoring, Stanford, CA, USA, 12–14 September 2023; pp. 963–969. [Google Scholar]
- Kalhori, H.; Alamdari, M.M.; Samali, B.; Halkon, B.; Rashidi, M. Ambient vibration dataset of a short-span cable-stayed bridge. Struct. Infrastruct. Eng. 2024, 1–16. [Google Scholar] [CrossRef]
- Cruz, F.R.G.; Garcia, R.G.; Manait, K.J.B.; Jastillano, W.J.D.; Guzman, D.J.L.; Manebo, M.B.; Payawal, J.M.G.; Uy, F.A.A. Structural health monitoring of footbridge using wireless MEMS accelerometers with fast fourrier transform. In Proceedings of the 2018 IEEE 10th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment and Management, HNICEM, Baguio City, Philippines, 29 November–2 December 2018. [Google Scholar] [CrossRef]
- Iban, N.; Soria, J.M.; Magdaleno, A.; Casado, C.; Diaz, I.M.; Lorenzana, A. Ad-hoc vibration monitoring system for a stress-ribbon footbridge: From design to operation. Smart Struct. Syst. 2018, 22, 13–25. [Google Scholar] [CrossRef]
- Komarizadehasl, S.; Lozano, F.; Lozano-Galant, J.A.; Ramos, G.; Turmo, J. Low-cost wireless structural health monitoring of bridges. Sensors 2022, 22, 5725. [Google Scholar] [CrossRef]
- Navabian, N.; Beskhyroun, S. Low-cost and high-performance wireless sensor nodes for structural health monitoring applications. In Proceedings of the European Workshop on Structural Health Monitoring, NDT.net, Manchester, UK, 10–13 July 2018. [Google Scholar]
- Sabato, A. Pedestrian bridge vibration monitoring using a wireless MEMS accelerometer board. In Proceedings of the 2015 IEEE 19th International Conference on Computer Supported Cooperative Work in Design, CSCWD, Calabria, Italy, 6–8 May 2015; pp. 437–442. [Google Scholar] [CrossRef]
- Abedin, S.; Biondi Vaccariello, A.M.; Cao, L.; Wu, R.; Yu, T.; Wang, X. Real time traffic monitoring of pedestrian bridge using distributed fiber optic sensing textile. In Proceedings of the SPIE-The International Society for Optical Engineering; Su, Z., Glisic, B., Limongelli, M.P., Eds.; SPIE: Bellingham, DC, USA, 2023; Volume 12486. [Google Scholar] [CrossRef]
- Abedin, S.; Biondi, A.M.; Wu, R.; Cao, L.; Wang, X. Structural health monitoring using a new type of distributed fiber optic smart textiles in combination with optical frequency domain reflectometry (OFDR): Taking a pedestrian bridge as case study. Sensors 2023, 23, 1591. [Google Scholar] [CrossRef]
- Biondi, A.M.; Guo, X.; Zhou, J.; Tang, Q.; Ghandi, H.; Goplan, B.; Hanna, T.; Ivey, J.; Yu, T.; Wang, X. Optical fiber sensing textile for temperature and strain distributed measurement. In Proceedings of the Nondestructive Characterization and Monitoring of Advanced Materials, Aerospace, Civil Infrastructure, and Transportation XV, Long Beach, CA, USA, 22–26 March 2021; Volume 11592, pp. 322–328. [Google Scholar] [CrossRef]
- Abedin, S.; Biondi, A.M.; Cao, L.; Wu, R.; Yu, T.; Wang, X. A textile embedded with distributed fiber optic sensors for pedestrian bridge monitoring. In Proceedings of the Experimental Vibration Analysis for Civil Engineering Structures; Springer: Cham, Switzerland, 2023; Volume 433, pp. 689–695. [Google Scholar] [CrossRef]
- Moschas, F.; Stiros, S. Dynamic deflections of a stifffootbridge using 100-Hz GNSS and accelerometer data. J. Surv. Eng. 2015, 141, 04015003. [Google Scholar] [CrossRef]
- Moschas, F.; Stiros, S. PLL bandwidth and noise in 100 Hz GPS measurements. GPS Solut. 2015, 19, 173–185. [Google Scholar] [CrossRef]
- Kumar, V.; Aygun, L.E.; Verma, N.; Sturm, J.C.; Glisic, B. Sensing sheets based on large area electronics for structural health monitoring of bridges. In Proceedings of the SPIE-The International Society for Optical Engineering; Lynch, J.P., Huang, H., Sohn, H., Wang, K.-W., Eds.; SPIE: Bellingham, DC, USA, 2019; Volume 10970. [Google Scholar] [CrossRef]
- Kumar, V.; Aygun, L.E.; Verma, N.; Sturm, J.C.; Glisic, B. Static and dynamic strain measurements using sensing sheets for damage detection. In Proceedings of the Congress of IABSE: 20th Congress of IABSE, New York, NY, USA, 20–22 September 2019; pp. 390–396. [Google Scholar]
- Aygun, L.E.; Kumar, V.; Weaver, C.; Gerber, M.; Wagner, S.; Verma, N.; Glisic, B.; Sturm, J.C. Large-area resistive strain sensing sheet for structural health monitoring. Sensors 2020, 20, 1386. [Google Scholar] [CrossRef] [PubMed]
- Glisic, B.; Verma, N. Very dense arrays of sensors for SHM based on large area electronics. In Structural Health Monitoring 2011: Condition-Based Maintenance and Intelligent Structures—Proceedings of the 8th International Workshop on Structural Health Monitoring; DEStech Publicatoins: Lancaster, PA, USA, 2011; Volume 2, pp. 1409–1416. [Google Scholar]
- Gentile, C.; Marrongelli, G. Measurement of human-induced vibrations on a lively footbridge by microwave remote sensing. In Proceedings of the 7th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Torino, Italy, 1–3 July 2015. [Google Scholar]
- Terrien, N.; Dhamene, F.; Zejli, H.; Habti, M.D. Structural health monitoring of a smart composite bridge using guided waves and acoustic emission techniques. In Proceedings of the 7th European Workshop on Structural Health Monitoring, EWSHM 2014-2nd European Conference of the Prognostics and Health Management (PHM) Society, Nantes, France, 8–11 July 2014; pp. 1139–1146. [Google Scholar]
- Hoskere, V.; Park, J.W.; Yoon, H.; Spencer, B.F. Vision-Based Modal Survey of Civil Infrastructure Using Unmanned Aerial Vehicles. J. Struct. Eng. 2019, 145, 04019062. [Google Scholar] [CrossRef]
- Fradelos, Y.; Thalla, O.; Biliani, I.; Stiros, S. Study of lateral displacements and the natural frequency of a pedestrian bridge using low-cost cameras. Sensors 2020, 20, 3217. [Google Scholar] [CrossRef] [PubMed]
- Brown, N.; Schumacher, T.; Vicente, M.A. Evaluation of a novel video- and laser-based displacement sensor prototype for civil infrastructure applications. J. Civ. Struct. Health Monit. 2021, 11, 265–281. [Google Scholar] [CrossRef]
- Ehrhart, M.; Lienhart, W. Development and evaluation of a long range image-based monitoring system for civil engineering structures. In Proceedings of the SPIE—The International Society for Optical Engineering; Shull, P.J., Ed.; SPIE: Bellingham, DC, USA, 2015; Volume 9437. [Google Scholar] [CrossRef]
- Ehrhart, M.; Lienhart, W. Image-based dynamic deformation monitoring of civil engineering structures from long ranges. In Proceedings of the SPIE-The International Society for Optical Engineering; Neil, K.S., Lam, E.Y., Eds.; SPIE: Bellingham, DC, USA, 2015; Volume 9405. [Google Scholar] [CrossRef]
- Shao, Y.; Li, L.; Li, J.; An, S.; Hao, H. Target-free 3D tiny structural vibration measurement based on deep learning and motion magnification. J. Sound Vib. 2022, 538, 117244. [Google Scholar] [CrossRef]
- V. Shajihan, S.A.; Hoang, T.; Mechitov, K.; Spencer, B.F. Wireless SmartVision system for synchronized displacement monitoring of railroad bridges. Comput. Civ. Infrastruct. Eng. 2022, 37, 1070–1088. [Google Scholar] [CrossRef]
- Walker, K.; Miller, T.H.; Gupta, R.; Shariati, A.; Schumacher, T. Development of virtual visual sensor applications for wood structural health monitoring. J. Test. Eval. 2018, 46, 24–32. [Google Scholar] [CrossRef]
- Wang, Y.; Lander, P.; Myung, H.; Song, W. Robotic sensing for assessing and monitoring civil infrastructures. In Sensor Technologies for Civil Infrastructures: Volume 2: Applications in Structural Health Monitoring; Elsevier: Amsterdam, The Netherlands, 2022. [Google Scholar]
- Heinemeyer, C.; Butz, C.; Keil, A.; Schlaich, M.; Goldbeck, A.; Trometor, S.; Lukic, M.; Chabrolin, B.; Lemaire, A.; Martin, P.O.; et al. Design of Lightweight Footbridges for Human Induced Vibrations – Background document in support to the implementation, harmonization and further development of the Eurocodes (JRC Scientific and Technical Reports, EUR 23984 EN). European Commission, Joint Research Centre, Institute for the Protection and Security of the Citizen; Publications Office of the European Union: Luxembourg, 2009; ISBN 9789279133879. ISSN 1018-5593. [Google Scholar] [CrossRef]
- Fu, Y.; Mechitov, K.; Hoang, T.; Kim, J.R.; Lee, D.H.; Spencer, B.F. Development and full-scale validation of high-fidelity data acquisition on a next-generation wireless smart sensor platform. Adv. Struct. Eng. 2019, 22, 3512–3533. [Google Scholar] [CrossRef]
- Domaneschi, M.; Apostoliti, C.; Cimellaro, G.P.; Glisic, B.; Kliewer, K. Monitoring footbridges using wireless mesh networks. In Proceedings of the 9th International Conference on Bridge Maintenance, Safety and Management, Melbourne, VIC, Australia, 9–13 July 2018; pp. 909–922. [Google Scholar]
- Al-Nasser, H.; Al-Zuriqat, T.; Dragos, K.; Chillón Geck, C.; Smarsly, K. A deep-learning-based approach towards identifying combined faults in structural health monitoring. In Proceedings of the European Workshop on Structural Health Monitoring, NDT.net, Potsdam, Germany, 10–13 June 2024. [Google Scholar] [CrossRef]
- Delgado, E.; Torralba, V.; Komarizadehasl, S.; Ramos, G.; Turmo, J.; Lozano-Galant, J. Evaluation of low-cost adaptable reliable accelerometer (LARA) technology. In Bridge Maintenance, Safety, Management, Digitalization and Sustainability; CRC Press: Boca Raton, FL, USA, 2024; pp. 2287–2294. [Google Scholar] [CrossRef]
- Sofi, M.; Lumantarna, E.; Mendis, P.A.; Duffield, C.; Rajabifard, A. Assessment of a pedestrian bridge dynamics using interferometric radar system IBIS-FS. Procedia Eng. 2017, 188, 33–40. [Google Scholar] [CrossRef]
- Qin, Y.; Gu, M. Research on Data Acquisition and Signal Processing of Fault Diagnosis System for Electromechanical Equipment. In Proceedings of the 2012 National Conference on Information Technology and Computer Science, Lanzhou, China, 16–18 November 2012. [Google Scholar] [CrossRef]
- Iacussi, L.; Brambilla, M.; Chiariotti, P.; Manzoni, S.; Cigada, A. Exploring crowdsensing potentials for structural health monitoring applications: Challenges and opportunities. In Proceedings of the 14th International Workshop on Structural Health Monitoring, Stanford, CA, USA, 12–14 September 2023; pp. 1033–1040. [Google Scholar]
- Ozer, E.; Feng, M.Q. Biomechanically influenced mobile and participatory pedestrian data for bridge monitoring. Int. J. Distrib. Sens. Netw. 2017, 13, 1550147717705240. [Google Scholar] [CrossRef]
- Quqa, S.; Giordano, P.F.; Limongelli, M.P. Shared micromobility-driven modal identification of urban bridges. Autom. Constr. 2022, 134, 104048. [Google Scholar] [CrossRef]
- Peng, Z.; Li, J.; Hao, H.; Yang, N. Mobile crowdsensing framework for drive-by-based dense spatial-resolution bridge mode shape identification. Eng. Struct. 2023, 292, 116515. [Google Scholar] [CrossRef]
- Ozer, E.; Feng, M.Q. Citizen-engaged structural health monitoring using smartphone sensors. In Proceedings of the 10th International Workshop on Structural Health Monitoring, IWSHM, Stanford, CA, USA, 1–3 September 2015; pp. 145–152. [Google Scholar] [CrossRef]
- Ozer, E.; Feng, M.Q. Synthesizing spatiotemporally sparse smartphone sensor data for bridge modal identification. Smart Mater. Struct. 2016, 25, 085007. [Google Scholar] [CrossRef]
- Miskiewicz, M.; Pyrzowski, L.; Wilde, K. Structural health monitoring system for suspension footbridge. In Proceedings of the 2017 Baltic Geodetic Congress (Geomatics), BGC Geomatics, Gdansk, Poland, 22–25 June 2017; pp. 321–325. [Google Scholar] [CrossRef]
- Morris, I.; Abdel-Jaber, H.; Glisic, B. Ground penetrating radar as a monitoring tool: Mapping internal features with attribute analysis. In Proceedings of the SHMII 2017-8th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Proceedings, Brisbane, Australia, 5–8 December 2017; pp. 686–693. [Google Scholar]
- Morris, I.; Abdel-Jaber, H.; Glisic, B. Ground penetrating radar for material characterization in structural health monitoring applications. In Proceedings of the 11th International Workshop on Structural Health Monitoring, IWSHM, Stanford, CA, USA, 12–14 September 2017; pp. 1669–1676. [Google Scholar] [CrossRef]
- Peralta, P.; Smarsly, K. Sustainable structural health monitoring using e-waste and recycled materials. In Current Perspectives and New Directions in Mechanics, Modelling and Design of Structural Systems; CRC Press: Boca Raton, FL, USA, 2022; pp. 1803–1807. [Google Scholar] [CrossRef]
- Ehrhart, M.; Lienhart, W. Monitoring of civil engineering structures using a state-of-the-art image assisted total station. J. Appl. Geod. 2015, 9, 174–182. [Google Scholar] [CrossRef]
- Yu, J.; Yan, B.; Meng, X.; Shao, X.; Ye, H. Measurement of bridge dynamic responses using network-based real-time kinematic gnss technique. J. Surv. Eng. 2016, 142, 04015013. [Google Scholar] [CrossRef]
- Miśkiewicz, M.; Pyrzowski, Ł.; Sobczyk, B. Short and long term measurements in assessment of FRP composite footbridge behavior. Materials 2020, 13, 525. [Google Scholar] [CrossRef]
- Morris, I.; Abdel-Jaber, H.; Glisic, B. Quantitative attribute analyses with ground penetrating radar for infrastructure assessments and structural health monitoring. Sensors 2019, 19, 1637. [Google Scholar] [CrossRef]
- Pereira, M.; Glisic, B. A hybrid approach for prediction of long-term behavior of concrete structures. J. Civ. Struct. Health Monit. 2022, 12, 891–911. [Google Scholar] [CrossRef]
- Bhowmik, B.; Hazra, B.; O’byrne, M.; Ghosh, B.; Pakrashi, V. Damping estimation of a pedestrian footbridge – an enhanced frequency-domain automated approach. J. Vibroeng. 2021, 23, 14–25. [Google Scholar] [CrossRef]
- Ngeljaratan, L.; Moustafa, M.A.; Pekcan, G. A compressive sensing method for processing and improving vision-based target-tracking signals for structural health monitoring. Comput. Civ. Infrastruct. Eng. 2021, 36, 1203–1223. [Google Scholar] [CrossRef]
- Dragos, K.; Magalhães, F.; Manolis, G.D.; Smarsly, K. Cross-spectrum-based synchronization of structural health monitoring data. In Proceedings of the European Workshop on Structural Health Monitoring; Springer: Cham, Switzerland, 2023; Volume 254, pp. 927–936. [Google Scholar] [CrossRef]
- Ramadan, W.; Ozer, E. Modal analysis under jittering and kernel clock distribution: Single-output identification. Proc. Inst. Civ. Eng. Smart Infrastruct. Constr. 2024, 177, 224–235. [Google Scholar] [CrossRef]
- Zhang, Y.; Wang, X.; Xia, Y. Few-shot classification for sensor anomalies with limited samples. J. Infrastruct. Intell. Resil. 2024, 3, 100087. [Google Scholar] [CrossRef]
- Miskiewicz, M.; Pyrzowski, L.; Chroscielewski, J.; Wilde, K. Structural health monitoring of composite shell footbridge for its design validation. In Proceedings of the 2016 Baltic Geodetic Congress (Geomatics), BGC Geomatics, Gdansk, Poland, 2–4 June 2016; pp. 228–233. [Google Scholar] [CrossRef]
- Omidalizarandi, M.; Kargoll, B.; Paffenholz, J.A.; Neumann, I. Accurate vision-based displacement and vibration analysis of bridge structures by means of an image-assisted total station. Adv. Mech. Eng. 2018, 10, 1687814018780052. [Google Scholar] [CrossRef]
- Xu, Y.; Brownjohn, J.; Kong, D. A non-contact vision-based system for multipoint displacement monitoring in a cable-stayed footbridge. Struct. Control. Health Monit. 2018, 25, e2155. [Google Scholar] [CrossRef]
- Yu, J.; Fang, Z.; Meng, X.; Xie, Y.; Fan, Q. Measurement of quasi-static and dynamic displacements of footbridges using the composite instrument of a smartstation and an accelerometer: Case studies. Remote. Sens. 2020, 12, 2635. [Google Scholar] [CrossRef]
- Ma, Z.; Choi, J.; Yang, L.; Sohn, H. Structural displacement estimation using accelerometer and FMCW millimeter wave radar. Mech. Syst. Signal Process. 2023, 182, 109582. [Google Scholar] [CrossRef]
- He, L.; Castoro, C.; Aloisio, A.; Zhang, Z.; Marano, G.C.; Gregori, A.; Deng, C.; Briseghella, B. Dynamic assessment, FE modelling and parametric updating of a butterfly-arch stress-ribbon pedestrian bridge. Struct. Infrastruct. Eng. 2022, 18, 1064–1075. [Google Scholar] [CrossRef]
- Huang, J.P.; He, L.Q.; Xue, J.Q.; Zhou, S.N.; Briseghella, B.; Castoro, C.; Aloisio, A.; Marano, G.C. Dynamic assessment of a stress-ribbon CFST arch bridge with SHM and NDE. In Life-Cycle of Structures and Infrastructure Systems; CRC Press: Boca Raton, FL, USA, 2023; pp. 2762–2769. [Google Scholar] [CrossRef]
- Colmenares, D.; Ruhani, S.; Karoumi, R. On the dynamic characterization of a two span post-tensioned concrete footbridge. Lect. Notes Civ. Eng. 2024, 515, 316–330. [Google Scholar] [CrossRef]
- Li, Z.; Lan, Y.; Lin, W. Indirect frequency identification of footbridges with pedestrians using the contact-point response of shared scooters. J. Bridge Eng. 2024, 29, 04024036. [Google Scholar] [CrossRef]
- Lu, C.; Xia, J.; Wang, X.; Wen, K.; Wang, Q.; Liu, Q. Analysis of human-induced vibration response and TLD vibration reduction of high fundamental frequency aluminum alloy footbridge. Structures 2024, 70, 107579. [Google Scholar] [CrossRef]
- Bassoli, E.; Gambarelli, P.; Simonini, L.; Vincenzi, L.; Savoia, M. Dynamic monitoring of the Pasternak footbridge using MEMS-based sensing system. In Proceedings of the SHMII 2015-7th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Torino, Italy, 1–3 July 2015. [Google Scholar]
- Cunha, Á.; Caetano, E.; Moutinho, C.; Hu, W.H.; Cardoso, A. Continuous dynamic monitoring of lively footbridges. In Proceedings of the IABSE Conference, Geneva 2015: Structural Engineering: Providing Solutions to Global Challenges-Report, Geneva, Switzerland, 23–25 September 2015; pp. 1303–1310. [Google Scholar]
- Bai, Y.; Sezen, H.; Yilmaz, A.; Qin, R. Bridge vibration measurements using different camera placements and techniques of computer vision and deep learning. Adv. Bridge Eng. 2023, 4, 25. [Google Scholar] [CrossRef]
- Ponsi, F.; Buoli, E.; Varzaneh, G.E.; Bassoli, E.; Briseghella, B.; Vincenzi, L. Vision-based approach for the static and dynamic monitoring of bridges. Procedia Struct. Integr. 2024, 62, 946–954. [Google Scholar] [CrossRef]
- Gamino, A.L.; Hune, R.P.; Santos, R.R.; Bragança, C.S.C.d.; Bittencourt, T.N.; Carvalho, H.; Futai, M.M. Dynamic analysis of a pedestrian bridge using monitoring and computational techniques. Struct. Infrastruct. Eng. 2024, 1–19. [Google Scholar] [CrossRef]
- Ponsi, F.; Buoli, E.; Varzaneh, G.E.; Bassoli, E.; Ghirelli, G.; Briseghella, B.; Vincenzi, L. Dynamic monitoring of a steel footbridge based on computer vision techniques. In Proceedings of the 10th International Operational Modal Analysis Conference (IOMAC 2024). IOMAC 2024; Lecture Notes in Civil Engineering; Springer: Cham, Switzerland, 2024; Volume 515, pp. 292–303. ISBN 978-303161424-8. [Google Scholar] [CrossRef]
- Miskiewicz, M.; Pyrzowski, L.; Wilde, K. Dynamic response of the suspended on a single cable footbridge. In Proceedings of the Society for Experimental Mechanics Series; Pakzad, S., Ed.; Springer: New York, NY, USA, 2020; pp. 211–217. [Google Scholar] [CrossRef]
- Kromanis, R. Characterizing footbridge response from cyclist crossings with computer vision-based monitoring. In Proceedings of the Civil Structural Health Monitoring Workshop; Springer: Cham, Switzerland, 2021; Volume 156, pp. 83–95. [Google Scholar] [CrossRef]
- Kromanis, R.; Elias, S. Measuring dynamic response of the wilford suspension bridge with a vision-based measurement system: A case study. In Proceedings of the European Workshop on Structural Health Monitoring; Rizzo, P., Milazzo, A., Eds.; Springer: Singapore, 2023; Volume 253, pp. 280–289. [Google Scholar] [CrossRef]
- Xia, Q.; Wu, W.L.; Li, F.N.; Zhou, X.Q.; Xu, Y.L.; Xia, Y. Temperature behaviors of an arch bridge through integration of field monitoring and unified numerical simulation. Adv. Struct. Eng. 2022, 25, 3492–3509. [Google Scholar] [CrossRef]
- Do Cabo, C.T.; Valente, N.A.; Mao, Z. A comparative analysis of imaging processing techniques for non-invasive structural health monitoring. IFAC-PapersOnLine 2022, 55, 150–154. [Google Scholar] [CrossRef]
- Sadhu, A.; Dunphy, K.; Zerpa, C. Investigation of vibration data-based human load monitoring system. Struct. Health Monit. 2021, 20, 791–803. [Google Scholar] [CrossRef]
- Hassoun, K.; Karaki, J.; Mustapha, S.; Kassir, A.; Dawy, Z.; Abi-Rached, H. Crowd load prediction on pedestrian bridges using Fiber Bragg grating sensors. In Life Cycle Analysis and Assessment in Civil Engineering: Towards an Integrated Vision; CRC Press: Boca Raton, FL, USA, 2018; pp. 1263–1270. [Google Scholar]
- Mustapha, S.; Kassir, A.; Hassoun, K.; Dawy, Z.; Abi-Rached, H. Estimation of crowd flow and load on pedestrian bridges using machine learning with sensor fusion. Autom. Constr. 2020, 112, 103092. [Google Scholar] [CrossRef]
- Liang, D.; Gao, Q.; Guo, B.; Zhang, Z.; Li, Z. Mechanical performance assessment of self-anchored suspension footbridge based on integral configuration. J. Civ. Struct. Health Monit. 2023, 13, 1683–1693. [Google Scholar] [CrossRef]
- Casciati, F.; Casciati, S.; Faravelli, L.; Vece, M. Validation of a data-fusion based solution in view of the real-time monitoring of cable-stayed bridges. Procedia Eng. 2017, 199, 2288–2293. [Google Scholar] [CrossRef]
- Ngeljaratan, L.; Moustafa, M.A. Structural health monitoring and seismic response assessment of bridge structures using target-tracking digital image correlation. Eng. Struct. 2020, 213, 110551. [Google Scholar] [CrossRef]
- Barros, R.C.; Paiva, F.M. On the use of radar interferometry for the structural monitoring of bridges. In Proceedings of the International Conference On Integrity-Reliability-Failure, Lisbon, Portugal, 22–26 July 2018; pp. 1287–1300. [Google Scholar]
- Weil, M.; Weijtjens, W.; Devriendt, C. Particle filter for temperature compensation of FBG strain sensors for event segmentation. In Proceedings of the 11th European Workshop on Structural Health Monitoring, NDT.net, Potsdam, Germany, 10–13 June 2024. [Google Scholar] [CrossRef]
- Ponsi, F.; Castagnetti, C.; Bassoli, E.; Mancini, F.; Vincenzi, L. Multi-sensor and multi-frequency data fusion for structural health monitoring. Lect. Notes Civ. Eng. 2024, 515, 281–291. [Google Scholar] [CrossRef]
- Abasi, A.; Sadhu, A. Performance evaluation of blind modal identification in large-scale civil infrastructure. Infrastructures 2022, 7, 98. [Google Scholar] [CrossRef]
- Brownjohn, J.M.W.; Bocian, M.; Hester, D.; Quattrone, A.; Hudson, W.; Moore, D.; Goh, S.; Lim, M.S. Footbridge system identification using wireless inertial measurement units for force and response measurements. J. Sound Vib. 2016, 384, 339–355. [Google Scholar] [CrossRef]
- Górski, P.; Tatara, M.; Stankiewicz, B. Vibration serviceability of all-GFRP cable-stayed footbridge under various service excitations. Measurement 2021, 183, 109822. [Google Scholar] [CrossRef]
- Jiang, K.; Han, Q.; Du, X.; Ni, P. Structural dynamic response reconstruction and virtual sensing using a sequence to sequence modeling with attention mechanism. Autom. Constr. 2021, 131, 103895. [Google Scholar] [CrossRef]
- Sadhu, A.; Narasimhan, S.; Goldack, A. Decentralized modal identification of a pony truss pedestrian bridge using wireless sensors. J. Bridge Eng. 2014, 19, 04014013. [Google Scholar] [CrossRef]
- Banerjee, S.; Saravanan, T.J. Utilizing data-driven algorithms for blind modal parameter identification of structures from output-only video measurements. Structures 2024, 63, 106410. [Google Scholar] [CrossRef]
- Domaneschi, M.; Cimellaro, G.P.; Kliewer, K.; Glišić, B. Modal identification of a flexible footbridge using output-only methods. In Proceedings of the 8th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Brisbane, Australia, 5–8 December 2017; pp. 594–603. [Google Scholar]
- Wang, L.; Song, R.; Wu, Y. Structural health monitoring for lightweight footbridges and pedestrian vibration assessment: A case study. In Proceedings of the 8th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Brisbane, Australia, 5–8 December 2017; pp. 1643–1652. [Google Scholar]
- Ozer, E. Smartphone-based vibration monitoring studies of a pedestrian link bridge. In Proceedings of the 11th European Workshop on Structural Health Monitoring. NDT.net, Potsdam, Germany, 10–13 June 2024. [Google Scholar] [CrossRef]
- Feng, K.; Hester, D.; Taylor, S.; O’Higgins, C.; Ferguson, A.; Zhu, Z.; Zou, G.; Lydon, M.; Early, J. Experimental modal identification of a pedestrian bridge through drive-by monitoring integrated with shared-mobility vehicles. Dev. Built Environ. 2024, 20, 100562. [Google Scholar] [CrossRef]
- Hacıefendioğlu, K.; Kahya, V.; Limongelli, M.G.; Okur, F.Y.; Altunışık, A.C.; Aslan, T.; Pembeoğlu, S.; Duman, C.; Bostan, A.; Aleit, H. Applications of optical flow methods and computer vision in structural health monitoring for enhanced modal identification. Structures 2024, 69, 107414. [Google Scholar] [CrossRef]
- Qiao, H.; Guan, H.; Jabbour, A.; Zhu, Y. Synchronous LoRa sensor nodes for modal identification in footbridge vibration monitoring. J. Civ. Struct. Health Monit. 2024, 15, 697–716. [Google Scholar] [CrossRef]
- He, L.; Zhang, Z.; Marano, G.C.; Briseghella, B.; Tavani, A.; Gregori, A.; Deng, C. Vibration-based condition monitoring of a stress-ribbon pedestrian bridge. In Life-Cycle Civil Engineering: Innovation, Theory and Practice; Chen, A., Ruan, X., Frangopol, D.M., Eds.; CRC Press: Boca Raton, FL, USA, 2020; pp. 1471–1475. [Google Scholar] [CrossRef]
- Venuti, F.; Domaneschi, M.; Lizana, M.; Glisic, B. Finite element models of a benchmark footbridge. Appl. Sci. 2021, 11, 9024. [Google Scholar] [CrossRef]
- Catbas, F.N.; Cano, J.A.; Luleci, F.; Walters, L.C.; Michlowitz, R. On the generation of digital data and models from point clouds: Application to a pedestrian bridge structure. Infrastructures 2024, 9, 6. [Google Scholar] [CrossRef]
- Li, Q.; Du, X.; Ni, P.; Han, Q.; Xu, K.; Bai, Y. Improved hierarchical Bayesian modeling framework with arbitrary polynomial chaos for probabilistic model updating. Mech. Syst. Signal Process. 2024, 215, 111409. [Google Scholar] [CrossRef]
- Yin, T. A practical bayesian framework for structural model updating and prediction. ASCE-ASME J. Risk Uncertain. Eng. Syst. A Civ. Eng. 2022, 8, 04021073. [Google Scholar] [CrossRef]
- Ao, W.K.; Hester, D.; O’Higgins, C.; Brownjohn, J. Tracking long-term modal behaviour of a footbridge and identifying potential SHM approaches. J. Civ. Struct. Health Monit. 2024, 14, 1311–1337. [Google Scholar] [CrossRef]
- Lutton, R.E.M.; McFarland, B.J.; McKenna, E.; Ohiggins, C.; Magee, B.; Gyftaki, E.; Kearney, J. How theoretical modelling compares to real-life through SHM using FBG sensors—A case study. In Proceedings of the 9th European Workshop on Structural Health Monitoring, Manchester, UK, 10–13 July 2018. [Google Scholar]
- Bjorngrim, N.; Hagman, O.; Wang, X.A. Multivariate Screening of the Weather Effect on Timber Bridge Movements. BioResources 2016, 11, 8890–8899. [Google Scholar] [CrossRef]
- Komarizadehasl, S.; Huguenet, P.; Lozano, F.; Lozano-Galant, J.A.; Turmo, J. Operational and analytical modal analysis of a bridge using low-cost wireless arduino-based accelerometers. Sensors 2022, 22, 9808. [Google Scholar] [CrossRef]
- Sun, Q.; Yan, W.J.; Ren, W.X.; Liu, L.L. Application of transmissibility measurements to operational modal analysis of railway, highway, and pedestrian cable-stayed bridges. Measurement 2019, 148, 106880. [Google Scholar] [CrossRef]
- Fontan, M.; Guerineau, L. A new SHM strategy using in situ modal analysis processing to monitor civil engineering structures. In Proceedings of the International Conference on Structural Health Monitoring of Intelligent Infrastructure, Porto, Portugal, 30 June–2 July 2021; pp. 1511–1516. [Google Scholar]
- Abdel-Jaber, H.; Glisic, B. Monitoring of prestress losses using long-gauge fiber optic sensors. In Proceedings of the International Society for Optical Engineering; SPIE: Bellingham, DC, USA, 2017; Volume 10170. [Google Scholar] [CrossRef]
- Abdel-Jaber, H.; Glisic, B. Monitoring of long-term prestress losses in prestressed concrete structures using fiber optic sensors. Struct. Health Monit. 2019, 18, 254–269. [Google Scholar] [CrossRef]
- Yu, Q.Q.; Gu, X.L.; Zeng, Y.H.; Zhang, W.P. Flexural behavior of Corrosion-Damaged prestressed concrete beams. Eng. Struct. 2022, 272, 114985. [Google Scholar] [CrossRef]
- Jeon, C.H.; Lee, J.B.; Lon, S.; Shim, C.S. Equivalent material model of corroded prestressing steel strand. J. Mater. Res. Technol. 2019, 8, 2450–2460. [Google Scholar] [CrossRef]
- Franceschini, L.; Vecchi, F.; Tondolo, F.; Belletti, B.; Sánchez Montero, J. Mechanical behaviour of corroded strands under chloride attack: A new constitutive law. Constr. Build. Mater. 2022, 316, 125872. [Google Scholar] [CrossRef]
- Calò, M.; Ruggieri, S.; Buitrago, M.; Nettis, A.; Adam, J.M.; Uva, G. An ML-based framework for predicting prestressing force reduction in reinforced concrete box-girder bridges with unbonded tendons. Eng. Struct. 2025, 325, 119400. [Google Scholar] [CrossRef]
- Pereira, M.; Glisic, B. Long-term prediction of rheological effects in concrete structures using probabilistic neural networks. In Proceedings of the International Conference on Structural Health Monitoring of Intelligent Infrastructure, Porto, Portugal, 30 June–2 July 2021; pp. 921–928. [Google Scholar]
- Desbazeille, M.; Sprynski, N.S.; Cattin, V.; Billères, M.; Jouanet, L.; Vidal, A. Monitoring of bridges by using static and dynamic data from MEMS accelerometers. In Proceedings of the Civil Structural Health Monitoring Workshop; Springer: Cham, Switzerland, 2021; Volume 156, pp. 725–745. [Google Scholar] [CrossRef]
- Oh, B.K.; Park, H.S.; Glisic, B. Prediction of long-term strain in concrete structure using convolutional neural networks, air temperature and time stamp of measurements. Autom. Constr. 2021, 126, 103665. [Google Scholar] [CrossRef]
- Górski, P.; Stankiewicz, B.; Tatara, M. Structural evaluation of all-GFRP cable-stayed footbridge after 20 years of service life. Steel Compos. Struct. 2018, 29, 527–544. [Google Scholar] [CrossRef]
- Catbas, N.; Dong, C.Z.; Celik, O.; Khuc, T. A vision for vision-based technologies for bridge health monitoring. In Maintenance, Safety, Risk, Management and Life-Cycle Performance of Bridges, Proceedings of the Ninth International Conference on Bridge Maintenance, Safety and Management (IABMAS 2018); Powers, N., Frangopol, D.M., Al-Mahaidi, R., Caprani, C., Eds.; CRC Press: Boca Raton, FL, USA, 2018; pp. 54–62. [Google Scholar]
- Dong, C.Z.; Bas, S.; Necati Catbas, F. A completely non-contact recognition system for bridge unit influence line using portable cameras and computer vision. Smart Struct. Syst. 2019, 24, 617–630. [Google Scholar] [CrossRef]
- Dong, C.Z.; Catbas, F.N. Structural identification of bridges using computer vision techniques. In Advances in Engineering Materials, Structures and Systems: Innovations, Mechanics and Applications; CRC Press: Boca Raton, FL, USA, 2019; pp. 2096–2100. [Google Scholar] [CrossRef]
- Zhao, R.; Wu, Y.; Feng, Z. Research for pedestrian steel bridge design of neural network in structural model updating. Shock Vib. 2022, 2022, 1057422. [Google Scholar] [CrossRef]
- Seon Park, H.; Hong, T.; Lee, D.E.; Kwan Oh, B.; Glisic, B. Long-term structural response prediction models for concrete structures using weather data, fiber-optic sensing, and convolutional neural network. Expert Syst. Appl. 2022, 201, 117152. [Google Scholar] [CrossRef]
- Duarte, A.; Ortiz, A.R. Feasibility study of using low-cost measurement devices for system identification using bayesian approaches. In Proceedings of the Society for Experimental Mechanics Series; Mao, Z., Ed.; Springer: Berlin, Germany, 2020; pp. 201–208. [Google Scholar] [CrossRef]
- Liu, Z.Y.; Yang, J.H.; Lam, H.F.; Peng, L.X. A Markov chain Monte Carlo-based Bayesian framework for system identification and uncertainty estimation of full-scale structures. Eng. Struct. 2023, 295, 116886. [Google Scholar] [CrossRef]
- Lim, J.; Yoon, H. Vision-based Input-Output System identification for pedestrian suspension bridges. Smart Struct. Syst. 2022, 29, 715–728. [Google Scholar] [CrossRef]
- Fábry, M.; Ároch, R.; Lipták, I. Application of system identification on a steel footbridge across the little Danube River. In Advances and Trends in Engineering Sciences and Technologies II: Proceedings of the 2nd International Conference on Engineering Sciences and Technologies, 29 June–1 July 2016, High Tatras Mountains, Tatranské Matliare, Slovak Republic; Al Ali, M., Platko, P., Eds.; CRC Press: Boca Raton, FL, USA, 2016; pp. 69–74. [Google Scholar]
- Donnelly, R.; Nogal, M.; Gorman, P.; O’Connor, A.J. A dynamic system identification of the ha’penny footbridge. In Management and Life-Cycle Performance of Bridges-Proceedings of the 9th International Conference on Bridge Maintenance, Safety and Management; Powers, N., Frangopol, D.M., Al-Mahaidi, R., Caprani, C., Eds.; CRC Press: Boca Raton, FL, USA, 2018; pp. 660–667. [Google Scholar]
- Manconi, M.; Moonen, S.F. A full computational framework for the temperature analysis of a flax-reinforced composite footbridge with field monitoring validation. Results Eng. 2024, 23, 102718. [Google Scholar] [CrossRef]
- Pereira, M.; Glisic, B. Detection and quantification of temperature sensor drift using probabilistic neural networks. Expert Syst. Appl. 2023, 213, 118884. [Google Scholar] [CrossRef]
- Omidalizarandi, M.; Herrmann, R.; Kargoll, B.; Marx, S.; Paffenholz, J.A.; Neumann, I. A validated robust and automatic procedure for vibration analysis of bridge structures using MEMS accelerometers. J. Appl. Geod. 2020, 14, 327–354. [Google Scholar] [CrossRef]
- Casciati, S.; Casciati, F.; Faravelli, L.; Bortoluzzi, D. Modelling the human induced vibrations in a cable-stayed pedestrian timber bridge. In Proceedings of the European Workshop on Structural Health Monitoring, Nantes, France, 8–11 July 2014; pp. 395–402. [Google Scholar]
- Kadota, T.; Miyamori, Y.; Watasaki, R.; Mikami, S.; Saito, T. Structural vibration characteristics of a pedestrian bridge by the 3D acceleration sensing of smart sensors and detailed 3D-FEM model. In Proceedings of the IABSE Conference. International Association for Bridge and Structural Engineering (IABSE): Elegance in Structures-Report, Nara, Japan, 13–15 May 2015; pp. 308–309. [Google Scholar]
- Ngan, J.W.; Caprani, C.C.; Bai, Y. Numerical validation framework of GFRP floor models using structural health monitoring. In Proceedings of the Australasian Conference on Mechanics of Structures and Materials, Brisbane, Australia, 4–7 December 2020; Volume 37, pp. 953–962. [Google Scholar] [CrossRef]
- Wang, M.; Kei Ao, W.; Bownjohn, J.; Xu, F. Completely non-contact modal testing of full-scale bridge in challenging conditions using vision sensing systems. Eng. Struct. 2022, 272, 114994. [Google Scholar] [CrossRef]
- Gres, S.; Ulriksen, M.D.; Döhler, M.; Johansen, R.J.; Andersen, P.; Damkilde, L.; Nielsen, S.A. Statistical methods for damage detection applied to civil structures. Procedia Eng. 2017, 199, 1919–1924. [Google Scholar] [CrossRef]
- Jimenez-Alonso, J.F.; Lopez-Martinez, J.; Blanco-Claraco, J.L.; Gonzalez-Diaz, R.; Saez, A. A topological entropy-based approach for damage detection of civil engineering structures. In Proceedings of the International Conference On Mechanical Models in Structural Engineering, Raspeig, Spain, 23–25 October 2019; pp. 35–42. [Google Scholar]
- Yin, T.; Zhu, H.P. An efficient algorithm for architecture design of Bayesian neural network in structural model updating. Comput. Civ. Infrastruct. Eng. 2020, 35, 354–372. [Google Scholar] [CrossRef]
- Ngoc-Nguyen, L.; Ngoc-Tran, H.; Khatir, S.; Le-Xuan, T.; Huu-Nguyen, Q.; De Roeck, G.; Bui-Tien, T.; Abdel Wahab, M. Damage assessment of suspension footbridge using vibration measurement data combined with a hybrid bee-genetic algorithm. Sci. Rep. 2022, 12, 20143. [Google Scholar] [CrossRef]
- You, T.; Gardoni, P.; Hurlebaus, S. Iterative damage index method for structural health monitoring. Struct. Monit. Maint. 2014, 1, 89–110. [Google Scholar] [CrossRef]
- Ellenberg, A.; Branco, L.; Krick, A.; Bartoli, I.; Kontsos, A. Use of unmanned aerial vehicle for quantitative infrastructure evaluation. J. Infrastruct. Syst. 2015, 21, 04014054. [Google Scholar] [CrossRef]
- Ghasr, M.T.; Barker, S.; Liu, C.; Chen, G.; Zoughi, R. Detection of delamination and effect of rebar corrosion in a pedestrian bridge deck using microwave sar imaging approach. In Proceedings of the 9th International Conference on Structural Health Monitoring of Intelligent Infrastructure: Transferring Research into Practice, St. Louis, MO, USA, 4–7 August 2019; pp. 1206–1211. [Google Scholar]
- Kumar, V.; Acot, B.; Aygun, L.E.; Wagner, S.; Verma, N.; Sturm, J.; Glisic, B. Detecting, localizing, and quantifying damage using two-dimensional sensing sheet: Lab test and field application. J. Civ. Struct. Health Monit. 2021, 11, 1055–1075. [Google Scholar] [CrossRef]
- Abadía, J.J.P.; Fritz, H.; Dadoulis, G.; Dragos, K.; Smarsly, K. Automated decision making in structural health monitoring using explainable artificial intelligence. In Proceedings of the Workshop on Intelligent Computing in Engineering: EG-ICE 2021 Workshop on Intelligent Computing in Engineering, Proceedings; Abualdenien, J., Borrmann, A., Ungureanu, L.-C., Hartmann, T., Eds.; Technische Universitat Berlin: Berlin, Germany, 2021; pp. 432–441. [Google Scholar]
- Kostic, B.; Gül, M. Vibration-based damage detection of bridges under varying temperature effects using time-series analysis and artificial neural networks. J. Bridge Eng. 2017, 22, 04017065. [Google Scholar] [CrossRef]
- Long, N.N.; Quyet, N.H.; Tung, N.X.; Thanh, B.T.; Hoa, T.N. Damage identification of suspension footbridge structures using new hunting-based algorithms. Eng. Technol. Appl. Sci. Res. 2023, 13, 11085–11090. [Google Scholar] [CrossRef]
- Ngoc-Nguyen, L.; Khatir, S.; Nguyen, H.Q.; Nguyen-Tran, H.; Bui-Ngoc, D.; Wahab, M.A.; Bui-Tien, T. Application of slime mould optimization algorithm on structural damage identification of suspension footbridge. In Recent Advances in Structural Health Monitoring and Engineering Structures: Select Proceedings of SHM and ES 2022; Springer Nature: Singapore, 2023; pp. 405–415. [Google Scholar] [CrossRef]
- He, Y.; Yin, T.; Wang, X. Design of multi-hidden-layer Bayesian neural networks for model updating. Adv. Struct. Eng. 2024, 28, 952–972. [Google Scholar] [CrossRef]
- Kostic, B.; Gül, M. A framework for damage detection under varying temperature effects using artificial neural networks and time series analysis. In Proceedings of the 10th International Workshop on Structural Health Monitoring, Stanford, CA, USA, 1–3 September 2015; pp. 759–766. [Google Scholar] [CrossRef]
- Dong, C.Z.; Celik, O.; Catbas, F.N.; O’Brien, E.J.; Taylor, S. Structural displacement monitoring using deep learning-based full field optical flow methods. Struct. Infrastruct. Eng. 2020, 16, 51–71. [Google Scholar] [CrossRef]
- Wootton, A.J.; Day, C.R.; Haycock, P.W. Echo State Network applications in structural health monitoring. In Proceedings of the 53rd Annual Conference of the British Institute of Non-Destructive Testing, Manchester, UK, 9–11 September 2014. [Google Scholar]
- Wootton, A.J.; Day, C.R.; Haycock, P.W. An Echo State Network approach to structural health monitoring. In Proceedings of the International Joint Conference on Neural Networks, Killarney, Ireland, 12–17 July 2015. [Google Scholar] [CrossRef]
- Wootton, A.J.; Butcher, J.B.; Kyriacou, T.; Day, C.R.; Haycock, P.W. Structural health monitoring of a footbridge using echo state networks and NARMAX. Eng. Appl. Artif. Intell. 2017, 64, 152–163. [Google Scholar] [CrossRef]
- Lee, J.; Hwang, C.; Park, J.; Park, T.R. (POSTER) enhancing structural health monitoring with on-device AI. In Proceedings of the 2024 20th International Conference on Distributed Computing in Smart Systems and the Internet of Things, Abu Dhabi, United Arab Emirates, 29 April–1 May 2024; pp. 762–764. [Google Scholar] [CrossRef]
- Li, Z.; Lan, Y.; Lin, W. Footbridge damage detection using smartphone-recorded responses of micromobility and convolutional neural networks. Autom. Constr. 2024, 166, 105587. [Google Scholar] [CrossRef]
- Zhao, S.; Tang, X.; Du, Y. Machine learning prediction and evaluation for structural damage comfort of suspension footbridge. Buildings 2024, 14, 1344. [Google Scholar] [CrossRef]
- Ahmadi, H.R.; Anvari, D. Health monitoring of pedestrian truss bridges using cone-shaped kernel distribution. Smart Struct. Syst. 2018, 22, 699–709. [Google Scholar] [CrossRef]
- Zhang, Z.; Sun, C. Structural damage identification via physics-guided machine learning: A methodology integrating pattern recognition with finite element model updating. Struct. Health Monit. 2021, 20, 1675–1688. [Google Scholar] [CrossRef]
- Sigurdardottir, D.H.; Glisic, B. On-site validation of fiber-optic methods for structural health monitoring: Streicker Bridge. J. Civ. Struct. Health Monit. 2015, 5, 529–549. [Google Scholar] [CrossRef]
- Yaghoubzadehfard, A.; Lumantarna, E.; Herath, N.; Sofi, M.; Rad, M. Ensemble learning-based structural health monitoring of a bridge using an interferometric radar system. J. Civ. Struct. Health Monit. 2024, 14, 1629–1650. [Google Scholar] [CrossRef]
- Liao, C.M.; Petryna, Y. Damage assessment of civil structures using wave propagation analysis and transmissibility functions. In Proceedings of the International Conference on Experimental Vibration Analysis for Civil Engineering Structures, Milan, Italy, 30 August–1 September 2023; Volume 224, pp. 3–13. [Google Scholar] [CrossRef]
- Pepi, C.; Gioffrè, M.; Grigoriu, M.D. Modeling uncertainties by data-driven bayesian updating for structural and damage detection. In Data Driven Methods for Civil Structural Health Monitoring and Resilience: Latest Developments and Applications; CRC Press: Boca Raton, FL, USA, 2023. [Google Scholar]
- Anastasopoulos, D.; Reynders, E.P.B.; François, S.; De Roeck, G.; Van Lysebetten, G.; Van Itterbeeck, P.; Huybrechts, N. Investigation of the influence of damage on modal strains of a fiber reinforced polymer footbridge using model updating for SHM purposes. In Proceedings of the International Conference on Structural Health Monitoring of Intelligent Infrastructure, Porto, Portugal, 30 June–2 July 2021; Volume 2021, pp. 571–579. [Google Scholar]
- Stiros, S.; Moschas, F. Rapid decay of a timber footbridge and changes in its modal frequencies derived from multiannual lateral deflection measurements. J. Bridge Eng. 2014, 19, 05014005. [Google Scholar] [CrossRef]
- Xi, P.S.; Ye, X.W.; Jin, T.; Chen, B. Structural performance monitoring of an urban footbridge. Struct. Monit. Maint. 2018, 5, 129–150. [Google Scholar] [CrossRef]
- Xia, Q.; Wu, W.; Li, F.; Xia, Y.; Ding, X.; Lam, W.H.K.; Chung, W.; Xu, Y. System design and demonstration of performance monitoring of a butterfly-shaped arch footbridge. Struct. Control. Health Monit. 2021, 28, e2738. [Google Scholar] [CrossRef]
- Nepomuceno, D.D.T.; Bennetts, J.; Webb, G.T.; Langhorne, M.; Johnson, M.; Macdonald, J.H.G.; Tryfonas, T.; Vardanega, P.J. Assessing the potential value of a SHM deployment on a proposed footbridge. In International Congress and Exhibition “Sustainable Civil Infrastructures: Innovative Infrastructure Geotechnology”; Springer: Cham, Switzerland, 2019; pp. 151–166. [Google Scholar] [CrossRef]
- Vardanega, P.J.; Webb, G.T.; Fidler, P.R.A.; Middleton, C.R. Assessing the potential value of bridge monitoring systems. Proc. Inst. Civ. Eng.-Bridge Eng. 2016, 169, 126–138. [Google Scholar] [CrossRef]
- Buffarini, G.; Clemente, P.; Giovinazzi, S.; Ormando, C.; Scafati, F. Structural assessment of the pedestrian bridge accessing Civita di Bagnoregio, Italy. J. Civ. Struct. Health Monit. 2023, 13, 1499–1516. [Google Scholar] [CrossRef]
- Kulpa, M.; Siwowski, T. Adaptation of the pipeline supporting structure into a footbridge using FRP deck with SHM. Procedia Struct. Integr. 2024, 64, 1339–1346. [Google Scholar] [CrossRef]
- Kulpa, M.; Siwowski, T. Field testing of the longest Polish footbridge made of FRP composites using distributed fibre optic sensor system. In Bridge Maintenance, Safety, Management, Digitalization and Sustainability, Proceedings of the 12th International Conference on Bridge Maintenance, Safety and Management, Copenhagen, Denmark, 24–28 June 2024; Jensen, J.S., Frangopol, D.M., Schmidt, J.W., Eds.; CRC Press: Boca Raton, FL, USA, 2024; pp. 929–936. [Google Scholar] [CrossRef]
- Marasco, G.; Moldovan, I.; Figueiredo, E.; Chiaia, B. Unsupervised transfer learning for structural health monitoring of urban pedestrian bridges. J. Civ. Struct. Health Monit. 2024, 14, 1487–1503. [Google Scholar] [CrossRef]
- Dragos, K.; Theiler, M.; Magalhães, F.; Moutinho, C.; Smarsly, K. On-board data synchronization in wireless structural health monitoring systems based on phase locking. Struct. Control. Health Monit. 2018, 25, e2248. [Google Scholar] [CrossRef]
- Jerath, K.; Brennan, S.; Lagoa, C. Bridging the gap between sensor noise modeling and sensor characterization. Measurement 2018, 116, 350–366. [Google Scholar] [CrossRef]
- Ding, X.; Shang, B.; Xie, C.; Xin, J.; Yu, F. Artificial intelligence in the COVID-19 pandemic: Balancing benefits and ethical challenges in China’s response. Humanit. Soc. Sci. Commun. 2025, 12, 245. [Google Scholar] [CrossRef]
- Nor, A.K.M.; Pedapati, S.R.; Muhammad, M.; Leiva, V. Abnormality Detection and Failure Prediction Using Explainable Bayesian Deep Learning: Methodology and Case Study with Industrial Data. Mathematics 2022, 10, 554. [Google Scholar] [CrossRef]
- Sarfarazi, S.; Mascolo, I.; Modano, M.; Guarracino, F. Application of Artificial Intelligence to Support Design and Analysis of Steel Structures. Metals 2025, 15, 408. [Google Scholar] [CrossRef]
- Huang, C.; Sun, X.; Zhang, Y. Tiny-Machine-Learning-Based Supply Canal Surface Condition Monitoring. Sensors 2024, 24, 4124. [Google Scholar] [CrossRef] [PubMed]
- Alzubi, A.; Megdadi, M.; Lin, D.; Reimann, J.; Calbaza, D.; MacDonald, R.; Yang, Z.; Alsaleem, F. In-MEMS analog neural network architecture for signal denoising. Sensors Actuators A Phys. 2025, 394, 116889. [Google Scholar] [CrossRef]
- Mousavi, V.; Rashidi, M.; Ghazimoghadam, S.; Samali, B. A data-driven digital twin model for bridge health monitoring using feature fusion and unsupervised deep learning. Inf. Fusion 2025, 126, 103534. [Google Scholar] [CrossRef]
- Mousavi, V.; Rashidi, M.; Ghazimoghadam, S.; Mohammadi, M.; Samali, B.; Devitt, J. Transformer-based time-series GAN for data augmentation in bridge digital twins. Autom. Constr. 2025, 175, 106208. [Google Scholar] [CrossRef]
- Guerrisi, G.; Bencivenni, G.; Schiavon, G.; Del Frate, F. On-Board Multispectral Image Compression with an Artificial Intelligence Based Algorithm. In Proceedings of the Dig Int Geosci Remote Sens Symp (IGARSS), Athens, Greece, 7–12 July 2024; pp. 2555–2559. [Google Scholar] [CrossRef]
- Beusen, B.; Ivashkovych, X.; Luyts, A.; Van Achteren, T. On-Board Hyperspectral Image Compression Using Vector-Quantized Auto Encoders. In Proceedings of the Dig Int Geosci Remote Sens Symp (IGARSS), Athens, Greece, 7–12 July 2024; pp. 1703–1707. [Google Scholar] [CrossRef]
- Boucetta, A.Y.; Baziz, M.; Hamdad, L.; Allal, I. Optimizing for Edge-AI Based Satellite Image Processing: A Survey of Techniques. In Proceedings of the 2024 IEEE Mediterranean and Middle-East Geoscience and Remote Sensing Symposium (M2GARSS), Oran, Algeria, 15–17 April 2024; pp. 83–87. [Google Scholar] [CrossRef]
- Gai, X. Application of flexible sensor multimodal data fusion system based on artificial synapse and machine learning in athletic injury prevention and health monitoring. Discov. Artif. Intell. 2025, 5, 31. [Google Scholar] [CrossRef]
- Zhang, H.; Upadhyay, D.; Zaman, M.; Jain, A.; Sampalli, S. SC-MLIDS: Fusion-based Machine Learning Framework for Intrusion Detection in Wireless Sensor Networks. Ad Hoc Netw. 2025, 175, 103871. [Google Scholar] [CrossRef]
- Fortune Business Insights. Smartphone Market Size, Share, Value|Growth Report [2029]. 2025. Available online: https://www.fortunebusinessinsights.com/industry-reports/smartphone-market-100308 (accessed on 1 March 2025).
- Sapkal, A.; Heisnam, L.; Kusi, S.S. Evolution of Cloud Computing: Milestones, Innovations, and Adoption Trends. Int. Res. J. Eng. Technol. 2024, 11, 548–563. [Google Scholar]
- Kaushal, P.; Mudhalwadkar, R.P. Internet of Things Enabled Electronic Tongue for Remote Monitoring of Water Quality. In Proceedings of the Smart Sensors Measurements and Instrumentation; Santhosh, K.V., Rao, K.G., Eds.; Springer: Singapore, 2021; pp. 71–79. [Google Scholar] [CrossRef]
- Preeti; Rana, C. IoT and Cloud Based health monitoring system Using Machine learning. In Proceedings of the P2022 5th International Conference on Contemporary Computing and Informatics (IC3I), Uttar Pradesh, India, 14–16 December 2022; pp. 1727–1732. [Google Scholar] [CrossRef]
- Jin, T.; Ye, X.W.; Que, W.M.; Wang, M.Y. Automatic detection, localization and quantification of structural cracks combining computer vision and crowd sensing technologies. Constr. Build. Mater. 2025, 476, 141150. [Google Scholar] [CrossRef]
- Komarizadehasl, S.; Mobaraki, B.; Ma, H.; Lozano-Galant, J.A.; Turmo, J. Development of a Low-Cost System for the Accurate Measurement of Structural Vibrations. Sensors 2021, 21, 6191. [Google Scholar] [CrossRef] [PubMed]
- Macchiarulo, V.; Milillo, P.; Blenkinsopp, C.; Giardina, G. Monitoring deformations of infrastructure networks: A fully automated GIS integration and analysis of InSAR time-series. Struct. Health Monit. 2022, 21, 1849–1878. [Google Scholar] [CrossRef]
- Calò, M.; Ruggieri, S.; Nettis, A.; Uva, G. A MTInSAR-Based Early Warning System to Appraise Deformations in Simply Supported Concrete Girder Bridges. Struct. Control. Health Monit. 2024, 2024, 8978782. [Google Scholar] [CrossRef]
- Calò, M.; Ruggieri, S.; Nettis, A.; Uva, G. A GIS Plugin for the Assessment of Deformations in Existing Bridge Portfolios via MTInSAR Data. Remote. Sens. 2024, 16, 4293. [Google Scholar] [CrossRef]
- Peduto, D.; Elia, F.; Montuori, R. Probabilistic analysis of settlement-induced damage to bridges in the city of Amsterdam (The Netherlands). Transp. Geotech. 2018, 14, 169–182. [Google Scholar] [CrossRef]
- Svedin, J.; Bernland, A.; Gustafsson, A.; Claar, E.; Luong, J. Small UAV-based SAR system using low-cost radar, position, and attitude sensors with onboard imaging capability. Int. J. Microw. Wirel. Technol. 2021, 13, 602–613. [Google Scholar] [CrossRef]
- Ojha, A.; Gupta, B. Evolving landscape of wireless sensor networks: A survey of trends, timelines, and future perspectives. Discov. Appl. Sci. 2025, 7, 825. [Google Scholar] [CrossRef]
- Fan, G.; Li, J.; Hao, H. Lost data recovery for structural health monitoring based on convolutional neural networks. Struct. Control. Health Monit. 2019, 26, e2433. [Google Scholar] [CrossRef]
- Fan, G.; Li, J.; Hao, H. Using deep learning technique for recovering the lost measurement data. Lect. Notes Civ. Eng. 2021, 101, 229–237. [Google Scholar] [CrossRef]
Figure 1.
The workflow used during the literature search.
Figure 2.
Historical trend of published papers in SHM of footbridges (2014–2024).
Figure 3.
Network of co-occuring keywords.
Figure 4.
The largest cluster within the connected network of co-authorship.
Figure 5.
The largest cluster within the connected network of countries.
Figure 6.
Major categories and their research purposes with colors separating the main categories. The number in each category shows the total papers for the respective category.
Figure 6.
Major categories and their research purposes with colors separating the main categories. The number in each category shows the total papers for the respective category.
Figure 7.
Energy-harvesting power [µW] obtained for different piezoelectric materials in terms of the considered Level of Service, (a) CPEH, (b) PPEH. Figure 6 pp. 710 Figure 7. pp 711 From: Energy harvesting from pedestrian-induced vibrations in footbridges with piezoelectric devices: A feasibility study, by Jiménez-Alonso et al. [19], Copyright 2023 by Imprint. Reproduced by permission of Taylor & Francis Group.
Figure 7.
Energy-harvesting power [µW] obtained for different piezoelectric materials in terms of the considered Level of Service, (a) CPEH, (b) PPEH. Figure 6 pp. 710 Figure 7. pp 711 From: Energy harvesting from pedestrian-induced vibrations in footbridges with piezoelectric devices: A feasibility study, by Jiménez-Alonso et al. [19], Copyright 2023 by Imprint. Reproduced by permission of Taylor & Francis Group.
Figure 8.
Sensor responses compared for (a) an individual jumping and (b) two people jumping. Reproduced with permission of Abedin et al. [35] CC-BY 4.0.
Figure 8.
Sensor responses compared for (a) an individual jumping and (b) two people jumping. Reproduced with permission of Abedin et al. [35] CC-BY 4.0.
Figure 9.
(a) Layers of how the sensing sheet is composed. (b) The prototype sensor and dimensions. (c) Expanded image showing single sensor. (d) Schematic diagram showing the design of the sensor. Used with permission of Springer Nature BV, from Detecting, localizing, and quantifying damage using two dimensional sensing sheet: lab test and field application, Aygun et al. [42], volume 11, copyright 2021; permission conveyed through Copyright Clearance Center, Inc.
Figure 9.
(a) Layers of how the sensing sheet is composed. (b) The prototype sensor and dimensions. (c) Expanded image showing single sensor. (d) Schematic diagram showing the design of the sensor. Used with permission of Springer Nature BV, from Detecting, localizing, and quantifying damage using two dimensional sensing sheet: lab test and field application, Aygun et al. [42], volume 11, copyright 2021; permission conveyed through Copyright Clearance Center, Inc.
Figure 10.
Illustration of the Brown et al. [48]’s deflection sensor: (a) fixed and (b) movable part. Used with permission of Springer Nature BV, from Evaluation of a novel video- and laser-based displacement sensor prototype for civil infrastructure applications, Brown et al. [48], volume 11, copyright 2011; permission conveyed through Copyright Clearance Center, Inc.
Figure 10.
Illustration of the Brown et al. [48]’s deflection sensor: (a) fixed and (b) movable part. Used with permission of Springer Nature BV, from Evaluation of a novel video- and laser-based displacement sensor prototype for civil infrastructure applications, Brown et al. [48], volume 11, copyright 2011; permission conveyed through Copyright Clearance Center, Inc.
Figure 11.
Functional diagram of the Xnode. Used with permission of John Wiley & Sons, from Wireless SmartVision system for synchronized displacement monitoring of railroad bridges, V. Shajihan et al. [52], volume 37, copyright 2022; permission conveyed through Copyright Clearance Center, Inc.
Figure 11.
Functional diagram of the Xnode. Used with permission of John Wiley & Sons, from Wireless SmartVision system for synchronized displacement monitoring of railroad bridges, V. Shajihan et al. [52], volume 37, copyright 2022; permission conveyed through Copyright Clearance Center, Inc.
Figure 12.
Historical graph showing vertical vibrations derived from various data acquisition: (a) NRTK-GNSS; (b) RTK-GNSS; (c) accelerometer. Used with permission of American Society of Civil Engineers, from Measurement of Bridge Dynamic Responses Using Network-Based Real-Time Kinematic GNSS Technique, Yu et al. [73], volume 142, copyright 2016; permission conveyed through Copyright Clearance Center, Inc.
Figure 12.
Historical graph showing vertical vibrations derived from various data acquisition: (a) NRTK-GNSS; (b) RTK-GNSS; (c) accelerometer. Used with permission of American Society of Civil Engineers, from Measurement of Bridge Dynamic Responses Using Network-Based Real-Time Kinematic GNSS Technique, Yu et al. [73], volume 142, copyright 2016; permission conveyed through Copyright Clearance Center, Inc.
Figure 13.
Graph showing the calculated x-axis and y-axis with the noise intensity in the collected data. Disparities of vibration magnitude in different years demonstrate the dissimilarity in agitations. Reproduced with permission of Fradelos et al. [47] CC-BY 4.0.
Figure 13.
Graph showing the calculated x-axis and y-axis with the noise intensity in the collected data. Disparities of vibration magnitude in different years demonstrate the dissimilarity in agitations. Reproduced with permission of Fradelos et al. [47] CC-BY 4.0.
Figure 14.
Overview of conventional deflection prediction with an FMCW: (a) the setup of the sensor; and (b) the general flow for deflection prediction. Used with permission of Elsevier Science & Technology Journals, from Structural displacement estimation using accelerometer and FMCW millimeter wave radar, Ma et al. [86], volume 182, copyright 2023; permission conveyed through Copyright Clearance Center, Inc.
Figure 14.
Overview of conventional deflection prediction with an FMCW: (a) the setup of the sensor; and (b) the general flow for deflection prediction. Used with permission of Elsevier Science & Technology Journals, from Structural displacement estimation using accelerometer and FMCW millimeter wave radar, Ma et al. [86], volume 182, copyright 2023; permission conveyed through Copyright Clearance Center, Inc.
Figure 15.
Ambient vertical accelerations (wind and traffic) (a), and walking in a closed loop (b). Reproduced with permission of Bassoli et al. [92].
Figure 15.
Ambient vertical accelerations (wind and traffic) (a), and walking in a closed loop (b). Reproduced with permission of Bassoli et al. [92].
Figure 16.
Accelerations data history of the deck acquired in computational analysis with (a) no TMDs; and (b) TMDs installed. Used with permission of Springer Nature BV, from Dynamic response of the suspended on a single cable footbridge, Miskiewicz et al. [98], volume 2, copyright 2020; permission conveyed through Copyright Clearance Center, Inc.
Figure 16.
Accelerations data history of the deck acquired in computational analysis with (a) no TMDs; and (b) TMDs installed. Used with permission of Springer Nature BV, from Dynamic response of the suspended on a single cable footbridge, Miskiewicz et al. [98], volume 2, copyright 2020; permission conveyed through Copyright Clearance Center, Inc.
Figure 17.
Sensor placement and data acquisition units on the pedestrian bridge. Reproduced with permission Xia et al. [101].
Figure 17.
Sensor placement and data acquisition units on the pedestrian bridge. Reproduced with permission Xia et al. [101].
Figure 18.
Ten persons—726 kg (top) fast, (bottom) slow. Used with permission of Taylor & Francis Group LLC-Books, from Crowd load prediction on pedestrian bridges using Fiber Bragg Grating sensors, Hassoun et al. [104], copyright 2018; permission conveyed through Copyright Clearance Center, Inc.
Figure 18.
Ten persons—726 kg (top) fast, (bottom) slow. Used with permission of Taylor & Francis Group LLC-Books, from Crowd load prediction on pedestrian bridges using Fiber Bragg Grating sensors, Hassoun et al. [104], copyright 2018; permission conveyed through Copyright Clearance Center, Inc.
Figure 19.
Dynamic deformation modes and the hypothetical estimation with interrelated eigenfrequencies and attenuation ratios obtained from the experiments triggered by jumping motion. Used with permission of Elsevier Science & Technology Journals, from Vibration serviceability of all-GFRP cable-stayed footbridge under various service excitations, Górski et al. [114], volume 13, copyright 2021; permission conveyed through Copyright Clearance Center, Inc.
Figure 19.
Dynamic deformation modes and the hypothetical estimation with interrelated eigenfrequencies and attenuation ratios obtained from the experiments triggered by jumping motion. Used with permission of Elsevier Science & Technology Journals, from Vibration serviceability of all-GFRP cable-stayed footbridge under various service excitations, Górski et al. [114], volume 13, copyright 2021; permission conveyed through Copyright Clearance Center, Inc.
Figure 20.
Flowchart of the proposed methodology. Used with permission of Elsevier Science & Technology Journals, from Utilizing data-driven algorithms for blind modal parameter identification of structures from output-only video measurements, Banerjee and Saravanan [117], volume 63, copyright 2024; permission conveyed through Copyright Clearance Center, Inc.
Figure 20.
Flowchart of the proposed methodology. Used with permission of Elsevier Science & Technology Journals, from Utilizing data-driven algorithms for blind modal parameter identification of structures from output-only video measurements, Banerjee and Saravanan [117], volume 63, copyright 2024; permission conveyed through Copyright Clearance Center, Inc.
Figure 21.
LARA: (a) enclosed sensors and measurements of LARA, (b) LARA’s printed circuit board. Adapted with permission of Komarizadehasl et al. [132] CC-BY 4.0.
Figure 21.
LARA: (a) enclosed sensors and measurements of LARA, (b) LARA’s printed circuit board. Adapted with permission of Komarizadehasl et al. [132] CC-BY 4.0.
Figure 22.
Overview of non-target displacement measurement method: (a) measurement locations; (b) extracted key-points as virtual makers; (c) matching key-points; (d) get dynamic displacements. Used with permission of Taylor & Francis Group LLC-Books, from A vision for vision-based technologies for bridge health monitoring, Catbas et al. [145], copyright 2018]; permission conveyed through Copyright Clearance Center, Inc.
Figure 22.
Overview of non-target displacement measurement method: (a) measurement locations; (b) extracted key-points as virtual makers; (c) matching key-points; (d) get dynamic displacements. Used with permission of Taylor & Francis Group LLC-Books, from A vision for vision-based technologies for bridge health monitoring, Catbas et al. [145], copyright 2018]; permission conveyed through Copyright Clearance Center, Inc.
Figure 23.
Structure of neural network: (a) backpropagation; (b) general regression neural network. Reproduced with permission of Zhao et al. [148] CC-BY 4.0.
Figure 23.
Structure of neural network: (a) backpropagation; (b) general regression neural network. Reproduced with permission of Zhao et al. [148] CC-BY 4.0.
Figure 24.
Overview of the explainable AI SHM approach. Reproduced with permission of Abadía et al. [170] CC-BY 4.0.
Figure 24.
Overview of the explainable AI SHM approach. Reproduced with permission of Abadía et al. [170] CC-BY 4.0.
Figure 25.
(a) IBIS-FS configuration. (b) Perfect distance ranging. (c) Side view of the footbridge. Reproduced with permission of Yaghoubzadehfard et al. [186] CC-BY 4.0.
Figure 25.
(a) IBIS-FS configuration. (b) Perfect distance ranging. (c) Side view of the footbridge. Reproduced with permission of Yaghoubzadehfard et al. [186] CC-BY 4.0.
Figure 26.
(a) Lateral view and measurements of the wooden footbridge. (b) Deck’s top view. (c,d) Various perspectives of the footbridge. Used with permission of American Society of Civil Engineers, from Rapid Decay of a Timber Footbridge and Changes in Its Modal Frequencies Derived from Multiannual Lateral Detection Measurements, Stiros and Moschas [190], volume 19, copyright 2014; permission conveyed through Copyright Clearance Center, Inc.
Figure 26.
(a) Lateral view and measurements of the wooden footbridge. (b) Deck’s top view. (c,d) Various perspectives of the footbridge. Used with permission of American Society of Civil Engineers, from Rapid Decay of a Timber Footbridge and Changes in Its Modal Frequencies Derived from Multiannual Lateral Detection Measurements, Stiros and Moschas [190], volume 19, copyright 2014; permission conveyed through Copyright Clearance Center, Inc.
Figure 27.
Historical trends in published papers on SHM of footbridges focusing on AI or utilizing AI as an analysis tool (2014–2024).
Figure 27.
Historical trends in published papers on SHM of footbridges focusing on AI or utilizing AI as an analysis tool (2014–2024).
Figure 28.
Annual number of publications returned by a Scopus keyword search for ‘InSAR’ from 2014 to 2024, with a red dotted line indicating the overall publication trend.
Figure 28.
Annual number of publications returned by a Scopus keyword search for ‘InSAR’ from 2014 to 2024, with a red dotted line indicating the overall publication trend.
Figure 29.
Annual number of publications returned by a Scopus keyword search for ‘MT-InSAR’ from 2014 to 2024, with a red dotted line indicating the overall publication trend.
Figure 29.
Annual number of publications returned by a Scopus keyword search for ‘MT-InSAR’ from 2014 to 2024, with a red dotted line indicating the overall publication trend.
Table 1.
Publications by type (2014–2024).
| Item Type | Number of Articles | Percentage of Total Included Publications |
|---|---|---|
| Books section | 3 | 1.8 |
| Conference paper | 76 | 44.4 |
| Journal article | 92 | 53.8 |
Table 2.
Top journals and their total publications (2014–2024).
| Journal Title | Number of Articles | Percentage of Total Included Publications |
|---|---|---|
| Journal of Civil Structural Health Monitoring | 10 | 5.85 |
| Sensors | 6 | 3.51 |
| Automation in Construction | 5 | 2.92 |
| Engineering Structures | 4 | 2.34 |
| Journal of Bridge Engineering | 4 | 2.34 |
Table 3.
Top conferences and their total publications (2014-2024).
| Conference Title | Number of Articles | Percentage of Total Included Publications |
|---|---|---|
| European Workshop on Structural Health Monitoring | 10 | 5.85 |
| International Conference on Structural Health Monitoring of Intelligent Infrastructure | 9 | 5.26 |
| International Workshop on Structural Health Monitoring | 6 | 3.51 |
| Proceedings of SPIE - The International Society for Optical Engineering | 5 | 2.92 |
| Bridge Safety, Maintenance, Management, Life-Cycle, Resilience and Sustainability | 3 | 1.75 |
Table 4.
Keywords list.
| Keyword | Occurrence | Average Year Published | Links | Total Link Strength | Percentage of Total Occurrence |
|---|---|---|---|---|---|
| structural health monitoring | 146 | 2020.2 | 32 | 486 | 22 |
| footbridges | 109 | 2020.1 | 32 | 394 | 16 |
| modal analysis | 32 | 2020.9 | 29 | 150 | 5 |
| damage detection | 30 | 2019.5 | 26 | 111 | 4 |
| fiber optic sensors | 24 | 2019.6 | 24 | 98 | 4 |
| monitoring | 21 | 2017.4 | 24 | 90 | 3 |
| health monitoring | 20 | 2021.5 | 24 | 74 | 3 |
| structural dynamics | 20 | 2019.9 | 22 | 87 | 3 |
| accelerometers | 17 | 2019.4 | 23 | 69 | 3 |
| vibration analysis | 17 | 2020.2 | 23 | 76 | 3 |
| natural frequencies | 15 | 2020.0 | 23 | 63 | 2 |
| fiber optics | 14 | 2020.6 | 20 | 59 | 2 |
| modal identification | 13 | 2020.0 | 19 | 52 | 2 |
| operational modal analysis | 13 | 2021.7 | 19 | 61 | 2 |
| structural analysis | 13 | 2018.5 | 19 | 50 | 2 |
| structural health | 13 | 2023.0 | 16 | 53 | 2 |
| structural health monitoring systems | 13 | 2020.9 | 17 | 51 | 2 |
| bridges | 12 | 2018.7 | 20 | 45 | 2 |
| finite element method | 12 | 2019.5 | 21 | 61 | 2 |
| footbridge | 11 | 2019.8 | 11 | 37 | 2 |
| computer vision | 10 | 2020.9 | 16 | 37 | 1 |
| concretes | 10 | 2020.8 | 19 | 37 | 1 |
| cameras | 9 | 2019.4 | 12 | 32 | 1 |
| neural networks | 9 | 2020.6 | 15 | 38 | 1 |
| uncertainty analysis | 9 | 2021.9 | 16 | 39 | 1 |
| wireless sensor networks | 9 | 2018.2 | 16 | 36 | 1 |
| bridge | 8 | 2019.6 | 15 | 33 | 1 |
| civil infrastructures | 8 | 2020.9 | 14 | 29 | 1 |
| frequency domain analysis | 8 | 2022.5 | 15 | 40 | 1 |
| life cycle | 8 | 2018.6 | 16 | 37 | 1 |
| machine learning | 8 | 2022.1 | 20 | 39 | 1 |
| modal parameters | 8 | 2022.2 | 19 | 48 | 1 |
| stochastic systems | 8 | 2021.5 | 19 | 44 | 1 |
Table 5.
Top 5 most productive authors (2014–2024).
| Author | Institution | Country | Papers | Percentage of Total Included Publications |
|---|---|---|---|---|
| Glisic, B. | Princeton University | USA | 20 | 11.70 |
| Ozer, E. | Columbia University | USA | 6 | 3.51 |
| Abdel-Jaber, H. | Princeton University | USA | 5 | 2.92 |
| Briseghella, B. | Fuzhou University | CHINA | 5 | 2.92 |
| Catbas, F.N. | University of Central Florida | USA | 5 | 2.92 |
Table 6.
List of the top 10 most productive countries (2014–2024).
| Country | Total Publications | Percentage of Total Included Publications |
|---|---|---|
| United States | 49 | 28.65 |
| China | 28 | 16.37 |
| Italy | 22 | 12.87 |
| United Kingdom | 13 | 7.60 |
| Poland | 10 | 5.85 |
| Germany | 8 | 4.68 |
| Spain | 8 | 4.68 |
| Hong Kong | 7 | 4.09 |
| South Korea | 7 | 4.09 |
| Ireland | 7 | 4.09 |
Table 7.
List of the top 10 most used sensors (2014–2024).
| Sensor Types | Number of Relevant Articles | Percentage of Total Included Publications |
|---|---|---|
| Accelerometer | 87 | 50.88 |
| Camera/imaging | 36 | 21.05 |
| Fiber optics | 27 | 15.79 |
| Temperature | 18 | 10.53 |
| Smartphone | 12 | 7.02 |
| GNSS/GPS | 7 | 4.09 |
| Radar | 7 | 4.09 |
| Strain gauge | 7 | 4.09 |
| Large area electronic sheet | 4 | 2.34 |
| Laser | 4 | 2.34 |
Table 8.
The number of sensors used.
| Number of Sensors Used | Number of Relevant Articles | Percentage of Total Included Publications |
|---|---|---|
| 1 sensor | 114 | 66.67 |
| 2 sensors | 32 | 18.71 |
| 3 sensors or more | 14 | 8.19 |
| Not specified | 11 | 6.43 |
Table 9.
Data Transmission Medium.
| Data Transmission | Number of Relevant Articles | Percentage of Total Included Publications |
|---|---|---|
| Not Specified | 125 | 73.10 |
| Wireless | 39 | 22.81 |
| Wired | 7 | 4.09 |
Note: 125 papers did not specify the data transmission medium; however, according to the SHM technology market report by MRA [15], wired systems presently hold a dominant position in the SHM market, and it is reasonable to assume that these papers likely employed a wired data transmission medium.
Table 10.
Major categories identified, sorted according to SHM workflow.
| Category | Papers | Percentage of Total Included Publications |
|---|---|---|
| Planning and workflow | 3 | 1.75 |
| Sensors | 34 | 19.88 |
| Data acquisition | 13 | 7.60 |
| Data processing and analysis | 86 | 50.29 |
| Damage detection | 27 | 15.79 |
| Reporting | 8 | 4.68 |
Table 11.
Accelerometer sensor development comparison.
| Publications | Wireless Network | Accelerometer | Sampling Rate (Hz) | Noise Density (g/√Hz) | Post Processing |
|---|---|---|---|---|---|
| Cruz et al. [29] | Wi-Fi | – | – | – | MATLAB |
| Iban et al. [30] | – | ADXL327 | 200 | 25 | – |
| Komarizadehasl et al. [31] | Cellular 4G | MPU9250 × 5 | 333 | 51 | MATLAB |
| Navabian and Beskhyroun [32] | Zigbee | ADXL355 | 100 | 25 | – |
| Sabato [33] | 2.4 Ghz ISM | SF1600SN.A | 30 | 0.3 | – |
Table 12.
Vision sensor comparison.
| Publications | Accuracy | MAC Value (Percent) |
|---|---|---|
| Brown et al. [48] | ±0.9 mm | 95 |
| Ehrhart and Lienhart [49] | a few millimeters | – |
| Ehrhart and Lienhart [50] | 0.2 mm | – |
| Hoskere et al. [46] | – | 99.6 |
| Shao et al. [51] | 0.167 pixel | – |
| V. Shajihan et al. [52] | 0.25 pixel | 99.31 |
| Walker et al. [53] | – | – |
| Wang et al. [54] | – | – |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Liew, J.; Rashidi, M.; Le, K.; Nazar, A.M.; Sorooshnia, E. Reviewing a Decade of Structural Health Monitoring in Footbridges: Advances, Challenges, and Future Directions. Remote Sens. 2025, 17, 2807. https://doi.org/10.3390/rs17162807
AMA Style
Liew J, Rashidi M, Le K, Nazar AM, Sorooshnia E. Reviewing a Decade of Structural Health Monitoring in Footbridges: Advances, Challenges, and Future Directions. Remote Sensing. 2025; 17(16):2807. https://doi.org/10.3390/rs17162807
Chicago/Turabian StyleLiew, JP, Maria Rashidi, Khoa Le, Ali Matin Nazar, and Ehsan Sorooshnia. 2025. "Reviewing a Decade of Structural Health Monitoring in Footbridges: Advances, Challenges, and Future Directions" Remote Sensing 17, no. 16: 2807. https://doi.org/10.3390/rs17162807
APA StyleLiew, J., Rashidi, M., Le, K., Nazar, A. M., & Sorooshnia, E. (2025). Reviewing a Decade of Structural Health Monitoring in Footbridges: Advances, Challenges, and Future Directions. Remote Sensing, 17(16), 2807. https://doi.org/10.3390/rs17162807
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
