<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>The NASA Digital VGH Program - Exploration of Methods and Final Results: Volume IV - B 747 Data 1978-1980: 1689 Hours</title>
      <link>https://trid.trb.org/View/2770080</link>
      <description><![CDATA[B 747 Digital Flight Data taken in 1978 through 1980 were analyzed to provide many statistical data useful for updating airliner airworthiness design criteria.]]></description>
      <pubDate>Sat, 05 Sep 2026 11:31:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2770080</guid>
    </item>
    <item>
      <title>The NASA Digital VGH Program - Exploration of Methods and Final Results: Volume V - DC 10 Data 1981-1982: 129 Hours</title>
      <link>https://trid.trb.org/View/2770143</link>
      <description><![CDATA[DC 10 Digital Flight Data Recorder data taken in 1981 through 1982 were analyzed to provide statistical data useful for updating airliner airworthiness design criteria.]]></description>
      <pubDate>Sat, 05 Sep 2026 11:31:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2770143</guid>
    </item>
    <item>
      <title>The NASA Digital VGH Program - Exploration of Methods and Final Results: Volume I - Development of Methods</title>
      <link>https://trid.trb.org/View/2769876</link>
      <description><![CDATA[Two hundred hours of Lockheed L 1011 Digital Flight Data Recorder data taken in 1973 were used to develop methods and procedures for obtaining statistical data useful for updating airliner airworthiness design criteria. Five thousand hours of additional data taken in 1978-1982 are reported in Volumes II, III, IV and V.]]></description>
      <pubDate>Sat, 05 Sep 2026 11:31:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2769876</guid>
    </item>
    <item>
      <title>The NASA Digital VGH Program - Exploration of Methods and Final Results: Volume II - L 1011 Data 1978-1979: 1619 Hours</title>
      <link>https://trid.trb.org/View/2769350</link>
      <description><![CDATA[L 1011 Digital Flight Data Recorder data taken in 1978-1979 were analyzed to provide many statistical data types useful for updating airliner airworthiness design criteria.]]></description>
      <pubDate>Sat, 05 Sep 2026 11:31:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2769350</guid>
    </item>
    <item>
      <title>The NASA Digital VGH Program - Exploration of Methods and Final Results: Volume III - B 727 Data 1978-1980: 1765 Hours</title>
      <link>https://trid.trb.org/View/2770123</link>
      <description><![CDATA[B 727 Digital Flight Data Recorder data taken in 1978 through 1980 were analyzed to provide many statistical data useful for updating airliner airworthiness design criteria.]]></description>
      <pubDate>Sat, 05 Sep 2026 11:31:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2770123</guid>
    </item>
    <item>
      <title>Towards Building-Level Modeling: Human-Driven Agentic Workflows for Multi-Source Data Synthesis</title>
      <link>https://trid.trb.org/View/2775061</link>
      <description><![CDATA[Travel demand modeling is limited by the aggregation problem: data are often synthesized at census-tract or traffic-zone scales that obscure building-level mobility patterns. This project develops a human-directed agentic workflow in which a transportation planner orchestrates specialized artificial intelligence (AI) agents that discover, evaluate, and integrate open geospatial data from sources such as OpenStreetMap, building footprint databases, census repositories, municipal portals, and selected sensor feeds. To address risks from plausible but incorrect agent outputs, the pilot will use a bounded proof-of-concept geography, source-provenance tracking, human review, conflict checks, and explicit accuracy metrics against benchmark data.

The work will demonstrate the workflow through an open Digital Forum where practitioners and students can access tools, data, and examples for building-level demand modeling. Expected outputs include a Human-AI Agentic Workflow methodology; an open-source software pipeline that converts enriched OpenStreetMap and related geospatial data into building-level travel demand inputs; a curated, version-controlled database of discoverable open-data sources with source-provenance records; validation procedures and accuracy metrics for handling conflicting sources; a pilot Digital Forum that serves as an educational resource and practitioner-facing decision-support tool; at least one peer-reviewed publication; and training or curriculum materials for students learning to manage artificial intelligence ecosystems in transportation planning.]]></description>
      <pubDate>Fri, 04 Sep 2026 15:25:32 GMT</pubDate>
      <guid>https://trid.trb.org/View/2775061</guid>
    </item>
    <item>
      <title>Analysis of National Travel Trends</title>
      <link>https://trid.trb.org/View/2775055</link>
      <description><![CDATA[Over the past few decades, the transportation landscape has undergone a significant transformation, driven by technological innovations, demographic shifts, and evolving socio-cultural norms. The proliferation of information and communication technologies has revolutionized how and where individuals undertake daily activities, increasingly substituting physical travel with virtual alternatives and reshaping work through telework. Changes in technology have also altered mode characteristics and performance and introduced new options such as micromobility, ride-hailing, autonomous vehicles, and robotic delivery.

Concurrently, demographic changes have altered travel needs and preferences. The population is aging, years spent in educational systems are increasing, and labor force participation rates have been declining. Migration trends from urban centers to suburban and exurban areas, as well as between metro areas, raise critical questions about service levels and infrastructure needs. Increasing diversity in lifestyles, attitudes, and values has added complexity to travel choices and patterns of activity, mobility, and time use.

This multi-stage project aims to shed light on trends in time, travel, transit, telework, and transportation spending over the past two decades and beyond. The effort will expand and extend ongoing analysis that the project team has conducted over many years, seeking to identify emerging trends such as behavior stabilization following the COVID-19 pandemic and the pace of influence from emerging travel options. Understanding these trends is crucial for addressing current and future challenges in transportation management, economic resilience, and societal well-being.

The project will utilize data from the American Time Use Survey (ATUS), the Consumer Expenditure Survey (CES), the American Community Survey (ACS), and the new National Household Travel Survey (NHTS 2025), supplemented by national metrics on e-commerce, telework participation, roadway vehicle miles traveled, transit ridership, and airline travel. An advanced data fusion approach will integrate these diverse datasets through compilation, cleansing, merging, and aggregation, applying rigorous statistical tools to ensure accuracy and reliability.]]></description>
      <pubDate>Fri, 04 Sep 2026 14:58:26 GMT</pubDate>
      <guid>https://trid.trb.org/View/2775055</guid>
    </item>
    <item>
      <title>Who Stays and Who Leaves? Examining Sample-Source Effect on Attrition in a Longitudinal Panel Survey</title>
      <link>https://trid.trb.org/View/2775054</link>
      <description><![CDATA[Longitudinal panel surveys are a cornerstone of travel behavior research, enabling measurement of behavioral change over time and facilitating before-and-after studies of policy interventions. Despite their scientific value, panel surveys suffer from a well-documented challenge: attrition. When respondents exit the panel non-randomly between different waves of longitudinal surveys, the resulting stayer sample becomes biased, thereby undermining both sample size and representativeness. While prior literature has examined how individual socio-demographic characteristics influence panel retention, little attention has been paid to whether the initial recruitment channel through which respondents enter a panel survey shapes their propensity to remain across multiple waves. This gap limits the ability of transportation researchers and survey practitioners to design longitudinal studies that yield reliable, representative data.

This project aims to examine whether recruitment strategy significantly influences panel survey retention, even after controlling for socio-economic, demographic, and attitudinal factors, and to identify which recruitment approach yields the highest retention rates. The study will use data from the COVID Future Survey, a three-wave nationwide longitudinal panel survey conducted between April 2020 and November 2021. Wave 1 recruited 8,385 valid respondents through three distinct channels: convenience sampling, mass email outreach, and a commercial online survey panel. Of the original sample, 22.5% responded to all three waves, 11.1% responded to Waves 1 and 2 only, and 66.4% responded only to Wave 1.

The project will formulate and estimate a Generalized Heterogeneous Data Model (GHDM) that jointly treats recruitment strategy membership and panel survey retention as endogenous outcomes, accounting for correlated latent attitudinal constructs and socio-demographic characteristics. Latent constructs to be incorporated include Risk Perception, Work-from-Home Propensity, and Virtual Activity Perception, specified through exploratory and confirmatory factor analyses to capture shared unobserved heterogeneity that simultaneously influences recruitment channel membership and panel retention propensity.

The insights derived from this project are expected to provide actionable guidance for transportation survey practitioners on recruitment strategy selection, improve the reliability and representativeness of longitudinal travel behavior data, and strengthen the data foundations underlying travel demand models and transportation planning decisions.]]></description>
      <pubDate>Fri, 04 Sep 2026 14:56:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2775054</guid>
    </item>
    <item>
      <title>Pavement performance prediction via a tabular foundation model</title>
      <link>https://trid.trb.org/View/2705452</link>
      <description><![CDATA[This study investigates the applicability of tabular foundation models to pavement performance prediction under small, imbalanced, and partially missing infrastructure datasets. Using Japan’s National Road Facility Inspection Database, we analyze 189 pavement segments along a 24.48 km section of National Route 6. The task is to forecast a three-level categorical performance condition state (Sound/Monitor/Repair) at the second inspection conducted five years after the first inspection, using inventory attributes and first-inspection records only. We apply TabPFN (Tabular Prior-Data Fitted Network) in a zero-shot manner (no dataset-specific training or hyperparameter tuning) under an order-based fold design that reflects an operational scenario where only a subset of segments is inspected and the remainder is inferred. With two folds, TabPFN achieves high predictive performance (Accuracy=0.926; Macro-F1=0.876) while maintaining strong detection of the minority class (Repair). In contrast, five supervised baselines, logistic regression, random forests, histogram-based gradient boosting, XGBoost, and CatBoost, show weaker minority-class detection, with Accuracy ranging from 0.878 to 0.905 and Macro-F1 from 0.606 to 0.841. We also assess the reliability of TabPFN’s predictive probabilities and observe that misclassifications concentrate in low-confidence ranges, highlighting their utility as uncertainty-aware outputs. In addition, based on the high predictive performance results of TabPFN, we discuss the practical potential for a 25% inspection workload reduction under a partial-inspection deployment scenario. Finally, a missing-data experiment reveals a notable degradation trend in predictive performance as the missing rate increases. These findings highlight both the practical potential and key deployment considerations of tabular foundation models for pavement asset management.]]></description>
      <pubDate>Thu, 03 Sep 2026 09:37:47 GMT</pubDate>
      <guid>https://trid.trb.org/View/2705452</guid>
    </item>
    <item>
      <title>Prototyping Automated Framework for Asset Extraction and Characterization from Mobile Lidar Data</title>
      <link>https://trid.trb.org/View/2752293</link>
      <description><![CDATA[To be able to handle the statewide mobile lidar data ODOT collects efficiently, there is a need for a robust workflow with significant automation that can extract many types of features by effectively leveraging various feature extraction tools and algorithms to support applications. In particular, road characterization has been identified as a key application for information extraction from mobile lidar data. In this literature review, the team first provides a brief overview of the model inventory of roadway elements (MIRE 2.0) to help identify the attributes that are of interest as well as feasible to extract from mobile lidar data. Next, a review of the existing commercial software summarizes the common challenges in using these solutions for feature extraction and characterization tasks. Lastly, the existing work related to road characterization is reviewed covering the topics of ground filtering and road extraction, cross slopes and grades, and horizontal curve, followed by a summary of the benefits of using mobile lidar data as well as the challenges of the existing studies.]]></description>
      <pubDate>Wed, 02 Sep 2026 09:20:46 GMT</pubDate>
      <guid>https://trid.trb.org/View/2752293</guid>
    </item>
    <item>
      <title>Spatiotemporal-Decoupled Training: Enhancing Car-Following Behavior Modeling With Cross-Spatiotemporal Generalization</title>
      <link>https://trid.trb.org/View/2672829</link>
      <description><![CDATA[This study explores the dynamics of a gated memory car-following system, with a focus on the challenges encountered when training models using fine-grained spatiotemporal data. To address the issues of redundant gradient updates and limited generalization inherent in traditional sequential training methods, a novel Spatiotemporal-Decoupled Training (SDT) method is proposed. This method enhances gradient variance by decoupling temporal dependencies and mixing trajectory segments from different vehicles, thereby improving model generalization performance and achieving a zero collision rate on test dataset. Experimental validation is carried out using three datasets (HighD, NGSIM-I80 and Lyft) and two basic models (GRU and LSTM) to assess the effectiveness of the proposed method. The results demonstrate significant improvements in model performance, including an 80% reduction in generalization error on the HighD dataset, a 14% reduction on the NGSIM-I80 dataset a 57% reduction on Lyft dataset, and the achievement of a Zero-collision rate on all test datasets, showcasing the potential of the SDT method for intelligent driving systems. Our code and experimental configurations are publicly available on GitHub to facilitate reproducibility and comparison: https://github.com/LiangzgJlu/Spatiotemporal-Decoupled-Training]]></description>
      <pubDate>Tue, 01 Sep 2026 14:02:31 GMT</pubDate>
      <guid>https://trid.trb.org/View/2672829</guid>
    </item>
    <item>
      <title>BELT-Fusion: Bayesian Evidential Late Fusion for Trustworthy V2X Perception</title>
      <link>https://trid.trb.org/View/2672827</link>
      <description><![CDATA[Vehicle-to-Everything (V2X) collaborative perception bolsters the performance of autonomous vehicles by overcoming occlusion challenges and expanding their sensing range. However, current methodologies frequently overlook the intrinsic uncertainties associated with localization inaccuracies, asynchronous measurements, and diverse agent models. Such uncertainties may weaken fusion reliability, leading to performance worse than single-vehicle perception. To address this pivotal challenge, we introduce BELT-Fusion, a cohesive probabilistic framework tailored for reliable V2X late fusion. Our framework offers two notable advantages. The first is explicit agent-level uncertainty modeling, where classification uncertainty is captured via evidential deep learning and regression uncertainty via Bayesian neural networks. This capability allows task-specific reliability assessments to be effortlessly incorporated into existing object detectors. Second, our framework introduces an uncertainty-aware adaptive fusion representation. This representation dynamically guides object selection and weight allocation based on measurable fusion-level uncertainty, ensuring reliable fusion results without retraining and enabling plug-and-play functionality. To validate BELT-Fusion’s efficacy, we conducted evaluations focusing on 3D object detection in both real-world and simulated scenarios using the OPV2V and DAIR-V2X datasets. BELT-Fusion improved AP@0.7 by 7.16% in noisy settings and 3.84% under ideal conditions over uncertainty-agnostic baselines, demonstrating its robustness under challenging and noisy conditions. Our code will be available at https://github.com/ZhiguoZhao/BELT-Fusion]]></description>
      <pubDate>Tue, 01 Sep 2026 14:02:31 GMT</pubDate>
      <guid>https://trid.trb.org/View/2672827</guid>
    </item>
    <item>
      <title>EAR-MM: An Efficient Adaptive and Robust Algorithm for Streaming Map Matching</title>
      <link>https://trid.trb.org/View/2672824</link>
      <description><![CDATA[Streaming map matching is essential for real-time location-based services, such as navigation systems and traffic monitoring, but maintaining accuracy under dynamic conditions with noisy GPS data presents significant challenges. We propose EAR-MM, an efficient, adaptive and robust streaming map matching algorithm designed to address these issues. To mitigate the impact of GPS noise, EAR-MM incorporates a candidate reusing mechanism that ensures stable candidate selection, enhancing robustness against noisy data. Additionally, a rollback mechanism is employed to leverage new data to correct previous matching errors, incrementally improving accuracy. EAR-MM also features an adaptive parameter tuning component, which dynamically adjusts parameters based on changing environmental conditions, ensuring consistent performance across diverse contexts. To meet real-time processing requirements, EAR-MM accelerates shortest-path calculations using a bidirectional Dijkstra algorithm with step-level caching, balancing memory consumption and computational efficiency effectively. Experimental evaluations on two real-world trajectory datasets from Chengdu and Wuxi demonstrate that EAR-MM significantly outperforms existing methods in terms of accuracy and efficiency, and adapts effectively to diverse geographic conditions. The source code is publicly available (https://github.com/Spatio-Temporal-Lab/StreamingTrajectoryMapMatching).]]></description>
      <pubDate>Tue, 01 Sep 2026 14:02:31 GMT</pubDate>
      <guid>https://trid.trb.org/View/2672824</guid>
    </item>
    <item>
      <title>Identification of Key Risk-Influencing Factors in Merchant Ship Collision Occurrence Using a Data-Driven Bayesian Network with Emphasis on Ship-Related Factors</title>
      <link>https://trid.trb.org/View/2737066</link>
      <description><![CDATA[Identifying key risk-influencing factors (RIFs) for collision occurrence is essential for preventing accidents. While most studies focus on accident severity, few address occurrence. Accident data alone are insufficient for understanding the occurrence, which requires comparing vessels involved in accidents with those that were not—necessitating exposure data. Automatic Identification System (AIS) data are often used for this purpose but limit the analysis to short-term or regional scopes. To overcome these limitations, this study constructed a ship × year-based database using long-term, global data from world fleet and maritime casualty databases. This study used a tree-augmented naïve Bayes-based Bayesian network (BN) to identify RIFs contributing to collision occurrence, focusing on ship-related RIFs. The relative importance of each factor was evaluated by introducing evidence into the BN and calculating the posterior probability of collision. The results reveal that the period and built year have a greater influence on collision occurrence than age. Other significant RIFs included the gross tonnage, draught, speed, and deadweight tonnage. Incorporating these RIFs into conventional risk assessments can enhance objectivity, reducing reliance on subjective expert judgement and strengthening decision-making in safety management.]]></description>
      <pubDate>Tue, 01 Sep 2026 14:02:29 GMT</pubDate>
      <guid>https://trid.trb.org/View/2737066</guid>
    </item>
    <item>
      <title>Enhancing ship recognition in satellite imagery: a YOLO-centric framework based on the updated ShipRSImageNet benchmark</title>
      <link>https://trid.trb.org/View/2743197</link>
      <description><![CDATA[To better monitor human maritime activities, this study presents an updated version of the ShipRSImageNet benchmark and evaluates recent fine-grained ship recognition models on high-resolution optical satellite imagery. The updated benchmark provides corrected and refined annotations with enhanced consistency. Approximately one-third of the samples (1206 out of 3435 images) have undergone correction, thereby improving label reliability for both model training and performance benchmarking. Based on this updated dataset, You Only Look Once (YOLO)-based models improved mean average precision (mAP) from the previously reported 0.66–0.70 to approximately 0.80, with improved precision and recall. Comparison experiments show that this gain mainly results from more reliable training supervision enabled by improved annotation quality. The normalized confusion matrices further indicate that the models can distinguish most ship categories under complex coastal backgrounds and diverse viewing geometries. Remaining errors are mainly caused by fine-grained class confusion, false positives in harbor environments, and missed detections of small vessels in crowded nearshore scenes. The updated ShipRSImageNet benchmark and ship recognition models developed in this study can support maritime activity monitoring, port vessel analysis, and non-cooperative vessel surveillance.]]></description>
      <pubDate>Tue, 01 Sep 2026 14:02:29 GMT</pubDate>
      <guid>https://trid.trb.org/View/2743197</guid>
    </item>
  </channel>
</rss>