<rss version="2.0" xmlns:atom="https://www.w3.org/2005/Atom">
  <channel>
    <title>Transport Research International Documentation (TRID)</title>
    <link>https://trid.trb.org/</link>
    <atom:link href="https://trid.trb.org/Record/RSS?s=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" rel="self" type="application/rss+xml" />
    <description></description>
    <language>en-us</language>
    <copyright>Copyright © 2026. National Academy of Sciences. All rights reserved.</copyright>
    <docs>http://blogs.law.harvard.edu/tech/rss</docs>
    <managingEditor>tris-trb@nas.edu (Bill McLeod)</managingEditor>
    <webMaster>tris-trb@nas.edu (Bill McLeod)</webMaster>
    <image>
      <title>Transport Research International Documentation (TRID)</title>
      <url>https://trid.trb.org/Images/PageHeader-wTitle.jpg</url>
      <link>https://trid.trb.org/</link>
    </image>
    <item>
      <title>Laboratory Study of IoT-Based Real-Time Culvert Blockage Detection</title>
      <link>https://trid.trb.org/View/2645426</link>
      <description><![CDATA[Extreme weather events are increasing the risk of flood-related infrastructure failures in areas where culverts play a critical role in water management. Blockages caused by sediment and debris can significantly reduce hydraulic capacity, leading to flooding and structural damage. Existing monitoring solutions such as visual inspection, conventional sensors, and closed-circuit television (CCTV)-based systems are often labor-intensive, costly, energy-demanding, resource-intensive, or insufficiently scalable for widespread application, especially in rural or resource-limited areas. This study presents a novel, low-cost Internet of Things (IoT)-enabled system for real-time culvert blockage detection that overcomes these limitations by eliminating reliance on visual data and complex computing infrastructure. By leveraging low-cost ultrasonic sensors and a geometry-independent decision-rule-based classification algorithm embedded in the microcontroller, the system demonstrated average detection accuracy of 87.0% (steady state) and 84.0% (surge) across 100 steady-state and 75 surge condition experiments in a controlled laboratory environment. The system successfully categorized blockage levels (0%, 25%, 50%, 75%, and 100%) and showed consistent performance under varying flow conditions. With a setup cost of approximately USD 35 per unit, the system offers high scalability, low power requirements, and practical deployable across large culvert networks, positioning it as a transformative tool for proactive culvert management and flood mitigation. Future work will focus on field trials under varied conditions for enhanced reliability, with potential applications in standardizing culvert monitoring practices and supporting proactive maintenance strategies.]]></description>
      <pubDate>Wed, 22 Apr 2026 16:15:29 GMT</pubDate>
      <guid>https://trid.trb.org/View/2645426</guid>
    </item>
    <item>
      <title>Multisource Data Fusion Approach for Predicting the Deterioration of Sign Structures along Highways</title>
      <link>https://trid.trb.org/View/2640613</link>
      <description><![CDATA[Sign structures along highways are essential components of the road transportation infrastructure system by providing useful information to drivers, but they are susceptible to fatigue cracking and corrosion. Accurately forecasting their deterioration enables transportation agencies to prioritize inspections and optimize maintenance budgets. In this paper, we develop a multisource data fusion approach based on autoencoder for predicting the deterioration of sign structures. The proposed model integrated multiple autoencoders to extract meaningful representations from raw data while the latent representation from each autoencoder corresponded to one head of the neural network and was independently processed on a multilayer perceptron (MLP). Furthermore, self-attention was incorporated into each head of the network to assign different weights, which allows the model to capture informative features and patterns within the data while downplaying the less relevant ones. In our experiment, structural attributes, traffic, and environmental data were collected from multisource databases to consider the various factors contributing to the fatigue cracking and corrosion of sign structures. Experimental results demonstrate that the proposed model achieves an AUC of 0.814 in predicting sign structure deterioration, outperforming the standard MLP (0.759), random forests (0.788), and XGBoost (0.794). Feature importance analysis using Shapley additive explanations (SHAP) identifies structure age and truck volume as the most influential predictors, with nine of the top fifteen features being environmental variables, underscoring the critical role of climatic stressors in structural degradation. By predicting sign structure deterioration, transportation agencies can implement proactive maintenance strategies and life cycle cost planning, ensuring the integrity and reliability of sign structures and enhancing road transportation safety.]]></description>
      <pubDate>Thu, 12 Mar 2026 14:02:12 GMT</pubDate>
      <guid>https://trid.trb.org/View/2640613</guid>
    </item>
    <item>
      <title>Risk Analysis Framework for Airport Infrastructure Systems: A Novel Integrated Approach</title>
      <link>https://trid.trb.org/View/2613629</link>
      <description><![CDATA[Airports are complex infrastructure systems characterized by diverse equipment, activities, and departments that must seamlessly integrate to fulfill their roles with minimal faults. Enhancing the reliability of airport systems is vital to maintain their operational efficiency. The purpose of this study is to develop a risk analysis framework based on integrated fuzzy-failure mode and effects analysis for assessing risks of airport systems. The potential failure modes associated with each system and failure causes are identified based on historical incidents and interviewing experts. The fuzzy approach is used to prioritize failure modes based on severity, occurrence, and detectability. The risk priority numbers of integrated fuzzy-failure mode and effects analysis (fuzzy-FMEA) are benchmarked with the traditional failure mode and effects analysis to highlight the benefits of considering uncertainties using fuzzy approach. The findings showed that the most critical failure modes are jet fuel tank rupture, runway approach light failure, weather station readings transmission failure, flickering approach lights, and instrument landing system glideslope deviation. The study indicates the appropriateness of the integrated fuzzy-FMEA approach for assessing risks of airport systems. It highlights that proactive maintenance practices, robust infrastructure, and continuous monitoring are essential and effective practices for ensuring the safety and reliability of airport operations.]]></description>
      <pubDate>Tue, 20 Jan 2026 10:17:48 GMT</pubDate>
      <guid>https://trid.trb.org/View/2613629</guid>
    </item>
    <item>
      <title>Understanding the Road Intervention Planning Process to Improve Efficiency and Effectiveness Using Digital Tools</title>
      <link>https://trid.trb.org/View/2613610</link>
      <description><![CDATA[Modern civilization relies heavily on the road infrastructure, which makes its management to avoid unnecessary disruptions to its use of paramount importance. Along with accelerating efforts in the use of data and algorithms for all types of management, there are also accelerating efforts to use data and algorithms to improve road infrastructure asset management. Many of these efforts are focused on improving specifically the subprocess of planning maintenance interventions, including the selection of time of execution and the work to be included. These efforts are, however, done in the absence of a clear view of how road maintenance interventions are planned in the real world, which in turn leads to the situation that research results are less impactful than they could be. To address this gap, this paper presents an example real-world road intervention planning process, which was developed based on existing documentation, and interviews with experts and managers of a road infrastructure organization (e.g., operator or agency). The individual tasks are described using a process diagram [business process modeling notation (BPMN) 2.0] and explanations as to when and who makes which decisions based on which information are given. Benefiting from the overview of the entire subprocess, this paper then discusses parts of the process that would benefit from targeted research on the methods and digital tools to improve it. This paper provides a starting point, a common understanding of the road intervention planning process, for researchers and developers to create tailored methods and digital tools to improve the planning process in terms of effectiveness and efficiency.]]></description>
      <pubDate>Tue, 20 Jan 2026 10:17:48 GMT</pubDate>
      <guid>https://trid.trb.org/View/2613610</guid>
    </item>
    <item>
      <title>Assessment of the Impacts of Climatic Factors and Infrastructure Characteristics on Gas Pipeline Failures</title>
      <link>https://trid.trb.org/View/2608088</link>
      <description><![CDATA[The increasing reliance on natural gas as a transitional energy source has underscored the importance of ensuring the safety and reliability of gas pipelines. This study examines failure patterns in gas transmission pipelines in the US, considering both infrastructure characteristics and climatic factors. Initial analyses of the spatial and temporal characteristics of pipeline incidents is performed based on the kernel density estimation (KDE) approach, and Moran’s 𝐼. Detailed analysis of the influence of various factors, including concurrent and antecedent climatic factors, on pipeline failures is achieved through negative binomial (NB) and random parameters negative binomial (RPNB) models, developed for both underground and aboveground pipelines. The RPNB model, which appears superior to the NB model for both underground and aboveground pipelines—as evidenced by the Akaike information criterion and Bayes information criterion—captures unobserved heterogeneity, enabling a more nuanced representation of complex, real-world dynamics. Marginal effect analysis based on the RPNB models provides a quantitative assessment of how specific factors influence pipeline incident probabilities. Precipitation and soil moisture emerged as the most influential climatic factors for underground pipeline failure, and precipitation was also found to be the primary factor affecting aboveground pipeline failure. Additionally, it was found that temperature-related factors potentially contributed to the failure of gas pipelines. The results provide useful insights regarding pipeline failure and controlling factors and will form the basis for additional detailed investigations and advancements in pipeline design, maintenance, and decision-making.]]></description>
      <pubDate>Mon, 29 Dec 2025 09:33:39 GMT</pubDate>
      <guid>https://trid.trb.org/View/2608088</guid>
    </item>
    <item>
      <title>Ex-Post Evaluation of the Effects of Simple Proactive Repairs with a Deterioration Prediction Model on Steel and Concrete Bridges Considering Sample Dropping Bias</title>
      <link>https://trid.trb.org/View/2607942</link>
      <description><![CDATA[During periodic inspections of bridges, the inspector’s first-aid works are sometimes implemented to improve safety and enhance preventative maintenance. It is important to evaluate the effectiveness of the inspector’s first-aid works, but since these works are implemented for damage with relatively high deterioration rates, inspection data samples with such damage that did not receive the inspector’s first-aid works cannot be obtained, which results in sample dropping bias in the data. In this study, this problem is solved by using a deterioration prediction model that takes into account the dropped samples. Then, the effect of the inspector’s first-aid works is quantitatively evaluated after incorporating a deterioration control effect of inhibiting the progression of deterioration by comparing the deterioration processes of when the inspector’s first-aid works are implemented and when they are not. Lastly, through an application case study using visual inspection data of an actual highway viaduct, the effect of the inspector’s first-aid works on inhibiting deterioration is quantitatively evaluated, and the timing of implementation of the inspector’s first-aid works is discussed.]]></description>
      <pubDate>Mon, 29 Dec 2025 09:33:39 GMT</pubDate>
      <guid>https://trid.trb.org/View/2607942</guid>
    </item>
    <item>
      <title>Enhancing Road Asset Management with CityGML Enriched by Public Inputs: A Comprehensive Approach to Pothole Repair Prioritization</title>
      <link>https://trid.trb.org/View/2607854</link>
      <description><![CDATA[AbstractInfrastructure asset management involves navigating complex sociotechnical challenges, requiring not only the technical assessment of physical assets but also the timely consideration of end-user satisfaction. This paper presents a novel approach to extracting and documenting asset repair prioritization decisions, emphasizing socioeconomic and demographic factors influencing those decisions. Pothole repair in the Toronto road network is used as a case example with a specific focus on pothole repair prioritization. Traditionally, pothole repairs have been prioritized primarily based on physical factors such as their size and location, and social considerations have been addressed in an unofficial/ad hoc manner, relying on the subjective judgment of decision makers rather than being systematically integrated into the decision-making process. This study proposes a systematic approach utilizing open GIS, specifically integrating technical and social aspects within the road network to uncover hidden patterns in past decisions, which can be applied to future scenarios. This approach is applied in the case study to distill collective knowledge and make informed decisions for prioritizing the potholes to be repaired. In the case study, an extended Geography Markup Language (CityGML) data model was used to link demographic attributes with potholes’ physical and functional characteristics. Statistical machine learning approaches were then applied to correlate such attributes with the priority of pothole repairs in the city of Toronto. To this end, pothole repair data in Toronto between the years 2017 and 2021 were used to train artificial neural networks, support vector machines, and random forest models. By incorporating demographic features, these machine learning models could estimate the urgency of repairing potholes with an accuracy of 74%. Therefore, by fusing physical and functional data with demographic information, the proposed method represents a significant step toward automating the decision process to systematically incorporate both subjective and objective aspects of decisions into a repair prioritization knowledge inference system.]]></description>
      <pubDate>Mon, 22 Dec 2025 16:07:14 GMT</pubDate>
      <guid>https://trid.trb.org/View/2607854</guid>
    </item>
    <item>
      <title>Mapping the Swiss Road and Rail Infrastructure Planning Process and Identifying Potential Areas for Improvement</title>
      <link>https://trid.trb.org/View/2594272</link>
      <description><![CDATA[The Swiss transportation infrastructure planning process, like all infrastructure planning processes, requires that the organizations involved determine how societal needs will change over time, how the current infrastructure may accommodate these needs and which interventions will best help modify the infrastructure if modification is required. This is challenging due to the divergence in needs of society, the many different organizations involved in the process, the long duration of the process and the iterative nature of the process. By addressing the efficiency of these processes, society’s changing needs will be accommodated more quickly. Efforts to improve infrastructure planning processes, however, lack an overarching view of the process and therefore cannot optimally improve the process. The work presented in this paper addresses this gap by modeling the Swiss road and rail infrastructure planning process for the first time, assessing its ability to meet societal needs, and making proposals for improvement. This is done by modeling the relevant portions of the planning process, analyzing how decisions are made within the process, and identifying challenges and opportunities in terms of efficiency and effectiveness. These challenges and opportunities are put in context using specific case studies in the canton of Zürich. The proposed opportunities for improvement are the adoption of an early planning network-benefit appraisal tool, the establishment of a coordinating body to align the many organizations involved, the systematic explicit consideration of the uncertainty related to planning decisions and the consideration for the time required for planning process tasks. It is suspected that these improvements would help policymakers shape planning processes to better enable planning objectives to be met efficiently and effectively.]]></description>
      <pubDate>Thu, 20 Nov 2025 17:06:25 GMT</pubDate>
      <guid>https://trid.trb.org/View/2594272</guid>
    </item>
    <item>
      <title>Risk-Based Spatial Safety Management System Approach: The Case of Trabzon Airport</title>
      <link>https://trid.trb.org/View/2594077</link>
      <description><![CDATA[Safety management systems are needed for the organized execution of complex and systematic activities in the aviation sector. Reducing the risks in the aviation industry to an acceptable level is possible with the success of the risk assessment process. Risk matrices are utilized in the assessment processes as stipulated in the international guidelines. This method, based on experts’ subjective evaluation, is frequently subject to criticism. This paper reviews the results of the actual risk assessment of airport assets through a case study. The study examined the usability of spatial analysis and statistical methods in the risk assessment process and questioned the significance levels of risk assessments. The risks identified by experts were significant at a 99% confidence level. The findings reveal that spatial and statistical analysis can support safety risk management activities. The proposed methodology aims to provide a geoinformatics perspective to airport risk management processes and increase the accuracy of risk assessment activities.]]></description>
      <pubDate>Tue, 18 Nov 2025 11:04:19 GMT</pubDate>
      <guid>https://trid.trb.org/View/2594077</guid>
    </item>
    <item>
      <title>Condition Evaluation of Mud Spots in Railroad Ballast Using Deep Learning</title>
      <link>https://trid.trb.org/View/2592025</link>
      <description><![CDATA[Railroad ballast fouling and degradation can lead to poor drainage, settlement, and reduced lateral stability. Reliable ballast condition evaluation during track inspections can help detect and eliminate such issues, which is essential to track stability and infrastructure sustainability and support safe transportation. Recently, computer vision–based ballast condition evaluation approaches have provided in-depth assessments of ballast, determining both the degradation level and aggregate size properties in an automated manner. However, current state-of-the-art approaches are less robust when analyzing ballast images collected from mud spots, in which aggregate particles are heavily covered with fine materials resulting from mud pumping. To provide more high-quality annotated mud spot data and augment the existing ballast image data set, this research proposes an innovative synthetic mud spot ballast image generation approach. To further enhance the capability of the ballast segmentation model and improve its zero-shot performance, a novel ballast particle segmentation model was developed based on the advanced Detection Transformer (DETR) and the foundation segmentation model, the Segment Anything Model (SAM). Evaluations were performed on the proposed synthetic ballast data set and the ballast segmentation model, which suggested that both methods contribute to better segmentation performance at mud spots. A field validation test using both proposed methods was conducted on a railroad segment in a revenue line in Indiana. The results showed that the proposed methods provide a more accurate and reliable evaluation of in-service ballast conditions with mud spots and can assist the sustainable management of railroad ballast infrastructure in practice.]]></description>
      <pubDate>Thu, 13 Nov 2025 16:59:49 GMT</pubDate>
      <guid>https://trid.trb.org/View/2592025</guid>
    </item>
    <item>
      <title>Optimizing Multicomponent Intervention Programs Considering Costs, Interdependencies, Possession Windows, and Service Loss Risks</title>
      <link>https://trid.trb.org/View/2561976</link>
      <description><![CDATA[Presently, upon identifying the necessary railway interventions for a given timeframe in the future, e.g., five years, railway managers frequently undertake an iterative process to sequence the required interventions. These iterations often happen from ten to two years ahead of execution. This sequencing requires extensive knowledge regarding the interventions, their costs, service impacts, their geolocation, construction procedures, availability of time and budget, and possibility of delaying interventions without significantly increasing the risk of failures. This paper proposes an algorithm aimed at improving efficiency and effectiveness of determining multicomponent railway intervention programs that results in the lowest service impacts. It considers the required possession windows, costs, and service impacts as well as interdependencies between interventions, service loss risks, and resources. A mixed integer linear programming model is developed to determine optimal intervention programs by maximizing the net benefit. The algorithm is demonstrated on a 25-km railway network in Switzerland. The results show up to 58% improvement in the net benefit of executing multicomponent intervention programs.]]></description>
      <pubDate>Wed, 01 Oct 2025 11:36:31 GMT</pubDate>
      <guid>https://trid.trb.org/View/2561976</guid>
    </item>
    <item>
      <title>Network-Level Infrastructure Asset Management with Multiagent Actor–Critic Reinforcement Learning: A Case of Highway Bridges</title>
      <link>https://trid.trb.org/View/2582432</link>
      <description><![CDATA[The network-level infrastructure asset management (NL-IAM) problem is excessively large and complex due to a large number of decision parameters, possible strategies, and underlying uncertainties. Although it often leads to far-from-reality strategies, simplifying the uncertainties has been a solution to reach acceptable strategies within feasible computational time. This study presents one of the first multiagent reinforcement learning based NL-IAM systems for achieving intervention strategies while considering different sources of uncertainties such as earthquakes, deterioration, and cost fluctuations. Alongside the added flexibility in the decision-making process, the results show that the proposed model significantly enhances stakeholders’ utilities by up to 33%, compared to the baseline, throughout the management horizon. Trained RL-based agents can inform asset managers with more desirable and closer-to-reality strategies leading to community-level enhancements in sustainability measures. This study also proposes effective approaches for handling the curse of dimensionality and reward engineering to pave the path for researchers toward enhancing NL-IAM frameworks.]]></description>
      <pubDate>Thu, 18 Sep 2025 17:02:11 GMT</pubDate>
      <guid>https://trid.trb.org/View/2582432</guid>
    </item>
    <item>
      <title>Signal Decomposition Techniques for Railroad Track Safety and Ride Quality Features Extraction: A Multicriteria Selection Framework</title>
      <link>https://trid.trb.org/View/2582419</link>
      <description><![CDATA[This paper investigates the feature extraction effectiveness of three signal decomposition techniques, short-time Fourier transform (STFT), Hilbert Huang transform (HHT), and wavelet transform (WT), for analyzing railroad track geometry data. Analysis of track geometry data is challenging due to the intrinsic properties of the data, such as their nonlinear and nonstationary nature. Capturing these properties can be quite important in assessing the dynamic performance of railway vehicles, which is highly sensitive to the nonlinear characteristics of track geometry, and which can lead to ride quality and safety issues. The study utilizes the profile (vertical alignment) signal from 145 km (90 mi) of mainline track on a major US Class 1 railroad. This track was segmented into approximately 1,800 segments of 80.5 m (264 ft) length and analyzed using the aforementioned methods. For each technique, performance metrics in the form of signal reconstruction error, noise sensitivity/robustness, time-frequency localization, and computation time were computed from the decomposition outputs. Furthermore, features including energy, time-frequency concentration, zero-crossing rate, kurtosis, and peak-to-peak amplitude were extracted using each method. These segments were classified into safety and ride quality categories to assess the discriminative capability of these features. Analysis of variance (F-test) and Fisher scores were evaluated for each feature to determine the performance efficacy of each decomposition technique. The results indicated that the wavelet transform provides best overall performance in terms of robustness, superior time-frequency localization, low computation time, and better discriminative power between safety and ride quality, making it highly effective for this application.]]></description>
      <pubDate>Thu, 18 Sep 2025 17:02:11 GMT</pubDate>
      <guid>https://trid.trb.org/View/2582419</guid>
    </item>
    <item>
      <title>Guiding State-Specific Studies on Highway Disasters: Examining Differences in Postdisaster Funding</title>
      <link>https://trid.trb.org/View/2563968</link>
      <description><![CDATA[The highway sector is frequently and severely impacted by natural disasters. To mitigate these effects and reduce the need for large relief budgets or emergency funds, it is crucial to study the resilience of the transportation network. As such, defining resilience in absolute terms is not enough; instead, it is equally important to analyze specific case studies and build a lessons-learned rapport. This is critical because it helps identify unique vulnerabilities and strengths that might be overlooked in broader analyses. However, current research on highway resilience often lacks empirical evidence to support the selection of these unique and significant case studies. This paper fills this knowledge gap by leveraging highway emergency relief funds (ERF) data as it provides insight into the resilience of a highway system. Consequently, a multistep methodology was adopted that included (1) standardizing and normalizing the ERF data; (2) a series of analysis of variance models to evaluate the influence of continuous variables on categorical ones; and (3) extracting unique data points causing variances followed by hierarchical clustering and descriptive statistics to pinpoint and analyze these critical points. Results revealed that while some states and disasters have had attention within the literature, others, such as Georgia and Idaho, share common highway funding urgencies due to flood disasters in 2019 that require more nuanced studies. To this effect, the detailed analysis of factors influencing postdisaster funding and resilience can offer case studies that lead to actionable insights to guide resource allocation, improve disaster response strategies, and enhance overall highway system resilience. This study contributes to the body of knowledge by identifying commonalities of states, years, and disasters that interact to affect time-related and monetary-related factors in the ERF for highways.]]></description>
      <pubDate>Tue, 16 Sep 2025 13:50:11 GMT</pubDate>
      <guid>https://trid.trb.org/View/2563968</guid>
    </item>
    <item>
      <title>Multilevel SVMD-AD-Prophet Model for Predicting the Track Irregularity of High-Speed Railways</title>
      <link>https://trid.trb.org/View/2576284</link>
      <description><![CDATA[Short-term prediction of track irregularity is essential for ensuring train operational safety and optimizing infrastructure maintenance costs. Existing prediction methods often overlook the impact of intermittent railway track maintenance, resulting in an inability to accurately forecast the evolving trends of track irregularities under maintenance influences. To address this limitation, this paper proposes a multilevel prediction approach that explicitly incorporates track maintenance into the characterization of track irregularity trends. The proposed method first utilizes the successive variational mode decomposition (SVMD) algorithm to decompose time-series track irregularity data into intrinsic mode functions (IMFs), isolating frequency components to improve short-term trend prediction. Subsequently, the development trend of each IMF is predicted using the Prophet model, augmented with an integrated time-series anomaly detection (AD) model to identify and account for maintenance dates. Furthermore, multiprediction error layers are designed to refine the model’s accuracy. The effectiveness and high precision of the proposed approach are validated through two real-world case studies based on field measurements of track irregularities. The results indicate that the proposed approach achieves minimal prediction deviation and outperforms several existing models in terms of accuracy. This approach offers valuable insights for early warning systems and timely maintenance strategies in railway track management.]]></description>
      <pubDate>Mon, 08 Sep 2025 14:54:43 GMT</pubDate>
      <guid>https://trid.trb.org/View/2576284</guid>
    </item>
  </channel>
</rss>