<?xml version="1.0" encoding="utf-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>TRB Publications Index</title><link>http://pubsindex.trb.org/</link><atom:link href="http://pubsindex.trb.org/common/TRIS Suite/feeds/rss.aspx?s=PHNlYXJjaD48cGFyYW1zPjxwYXJhbSBuYW1lPSJzcGVjaWZpY3Rlcm1zIiB2YWx1ZT0iOTI2wqE1MTA1NcKhNTEwODTCoTI5NznCoTExMDUwN8KhOTQ0NzXCoTUwNzDCoTU0MjQiIC8%2BPHBhcmFtIG5hbWU9ImxvY2F0aW9uIiB2YWx1ZT0iMiIgLz48cGFyYW0gbmFtZT0ic3ViamVjdGxvZ2ljIiB2YWx1ZT0ib3IiIC8%2BPHBhcmFtIG5hbWU9InRlcm1zbG9naWMiIHZhbHVlPSJvciIgLz48L3BhcmFtcz48ZmlsdGVycyAvPjxyYW5nZXMgLz48c29ydHM%2BPHNvcnQgZmllbGQ9InB1Ymxpc2hlZCIgb3JkZXI9ImRlc2MiIC8%2BPC9zb3J0cz48cGVyc2lzdHM%2BPHBlcnNpc3QgbmFtZT0icmFuZ2V0eXBlIiB2YWx1ZT0icHVibGlzaGVkZGF0ZSIgLz48L3BlcnNpc3RzPjwvc2VhcmNoPg%3D%3D" rel="self" type="application/rss+xml" /><description></description><language>en-us</language><copyright>Copyright © 2015. 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>TRB Publications Index</title><url>http://pubsindex.trb.org/Images/PageHeader-wTitle.png</url><link>http://pubsindex.trb.org/</link></image><item><title>Meta-Reinforcement Learning with Hypernetworks for Variable Speed Limit Control under Adverse Weather and Work Zones</title><link>http://pubsindex.trb.org/view/2675553</link><description><![CDATA[Variable speed limits (VSL) have been widely implemented to alleviate highway congestion and enhance operational efficiency. However, most existing studies focus on fixed traffic scenarios, making them inadequate when addressing uncertainties such as fluctuating traffic flows, extreme weather conditions, and construction-induced closures. Consequently, traditional VSL control strategies exhibit limited adaptability and generalization capability in unfamiliar scenarios. To overcome these limitations, this paper proposes a VSL control strategy based on Meta-Reinforcement Learning (Meta-RL) and Multi-Agent Proximal Policy Optimization (MAPPO) (Meta-MAPPO). This method leverages the meta-learning mechanism of Meta-RL and integrates a Hypernetwork module to dynamically adjust the network parameters of the control policy. By doing so, it adapts to diverse traffic scenarios and environmental disturbances, facilitating rapid policy transfer across scenarios and enhancing control performance. The training results demonstrate that Meta-MAPPO achieves faster convergence and superior model performance than MAPPO and Meta Multi-Agent Soft Actor-Critic (Meta-MASAC). Simulation experiments reveal that, compared with traditional feedback control methods and conventional multi-agent RL approaches, Meta-MAPPO exhibits significant advantages in unseen scenarios: it effectively mitigates traffic congestion and substantially reduces total travel time. The findings provide a more applicable solution for the practical implementation of VSL and offer valuable insights for further exploration of multi-agent methodologies in intelligent transportation systems.]]></description><pubDate>Mon, 02 Mar 2026 13:29:09 GMT</pubDate><guid>http://pubsindex.trb.org/view/2675553</guid></item><item><title>Factors Affecting Urban Sustainable Development by Metro Using the GCN–MGWR model</title><link>http://pubsindex.trb.org/view/2663302</link><description><![CDATA[This study investigates the relationship between urban subway passenger flow and land use intensity, proposing an innovative hybrid model that combines graph convolutional networks and multiscale geographically weighted regression (GMGWR). This model addresses the limitations of traditional methods in handling nonlinearity and spatial heterogeneity. Using metro data from Chengdu, Sichuan, China, this study analyzes the effects of various land use types on metro passenger flow during different time periods, revealing the spatial and temporal dynamics of land use on the urban rail transit system. The results indicate that land use characteristics are key determinants of urban rail transit passenger flow and that the effects of land use intensity on metro passenger flow exhibit dynamic characteristics that change with time and space. The innovation of this study lies in integrating machine learning and spatial econometrics methods. The proposed GMGWR model provides a more accurate representation of the complex nonlinear relationship between land use and metro passenger flow, offering urban transportation planners valuable strategies to enhance public transportation systems. By strategically planning land use around metro stations and promoting transit-oriented development policies, it is possible to create livable, pedestrian-friendly communities that foster green, sustainable urban growth.]]></description><pubDate>Wed, 04 Feb 2026 16:29:39 GMT</pubDate><guid>http://pubsindex.trb.org/view/2663302</guid></item><item><title>Treatment Effectiveness Evaluation of Mandatory Pre-Right-Turn Stops for Large Vehicles at Signalized Intersections: Combining Causal Inference with the Wasserstein Generative Adversarial Network with Gradient Penalty Method</title><link>http://pubsindex.trb.org/view/2663047</link><description><![CDATA[In China, more than 30% of intersection traffic crashes are related to right-hand turns, with large vehicle crashes consistently being the most prevalent. Novel countermeasures termed mandatory pre-right-turn stops for large vehicles at signalized intersections were widely implemented. Understanding their effects is critical for improving road safety. Traditional statistical approaches are affected by confounding bias, and because of the problem of overfitting, causal machine learning methods often fail to produce accurate results in small sample size . To overcome these challenges, a novel treatment effect evaluation system that combines a causal inference method with a data generation method is proposed in this paper. The Wasserstein generative adversarial network with gradient penalty (WGAN-GP) is used to generate additional control group data, thereby increasing the sample size. Findings demonstrate that the WGAN-GP-generated data closely match real data. The effects of countermeasures are evaluated using doubly robust estimation with neural networks (DR-NNs). This method effectively models nonlinear relationships between variables and addresses confounding bias. Both the combination of a right-turn stop sign for large vehicles and right-turn danger zones, and the right-turn stop sign alone, significantly reduce right-turn crashes, with crash modification factors (CMFs) of 0.53 and 0.56, respectively. The analysis of heterogeneous treatment effects indicates that the combination is less effective at skewed intersections, and intersections without bicycle lanes show higher CMFs for the right-turn stop sign alone. These results underscore the importance of bicycle lanes and highlight the need for improved implementation of right-turn danger zones at skewed intersections.]]></description><pubDate>Fri, 30 Jan 2026 09:03:53 GMT</pubDate><guid>http://pubsindex.trb.org/view/2663047</guid></item><item><title>Study of Ship Path Planning in Complex Environments Based on Spatiotemporal Reinforcement Learning</title><link>http://pubsindex.trb.org/view/2659377</link><description><![CDATA[To address the challenges in ship path planning and collision avoidance in complex marine environments, this study proposes a deep reinforcement learning (DRL)-based local path planning algorithm that integrates spatiotemporal feature modeling, aiming to achieve unified ship collision avoidance and motion control. Firstly, an end-to-end state space and action space framework is established, enabling the system to directly output control commands based on perceptual information, thereby bridging the entire pipeline from perception to control. Secondly, a network architecture incorporating spatial position encoding, temporal context modeling, and an attention mechanism is designed to enhance feature modeling and decision-making capabilities in complex environments. Finally, a reward function system combining navigation objectives and collision avoidance requirements is constructed, alleviating the sparse reward problem in reinforcement learning and accelerating training convergence. Comparative experiments conducted in various typical scenarios demonstrate the superiority of the proposed algorithm with reference to path efficiency, safety, and control stability, providing robust technical support for the autonomous navigation of intelligent ships.]]></description><pubDate>Tue, 27 Jan 2026 17:08:28 GMT</pubDate><guid>http://pubsindex.trb.org/view/2659377</guid></item><item><title>Estimating Bicycle Volume Levels at Urban Intersections using Large Language Models</title><link>http://pubsindex.trb.org/view/2659296</link><description><![CDATA[Bicycle volume estimation is essential for effective city planning and ensuring the safety of vulnerable road users. Traditional approaches often involve resource-intensive data collection, such as short-term counts expanded using continuous counts (CCs), or the development of direct-demand models from site-specific data such as geometry, socioeconomic, and land use. Although recent efforts have explored crowdsourced data, such sources may be biased and not consistently available across jurisdictions. This study explores the potential of large language models (LLMs), specifically ChatGPT-4o mini, as a low-cost and accessible alternative for estimating bicycle volume levels at intersections.A data set of CCs over 12 months and site characteristics for 158 signalized intersections in the Region of Waterloo, Ontario, Canada, was compiled. From this, the average annual daily bicycle volume was computed, and intersections were categorized into low, medium, or high volume levels.Eight LLMs and three benchmark approaches, including a Naïve (random), an ordered logit (OL) model, and human survey responses, were applied to the data set. Accuracy in predicting the correct volume category was the evaluation metric. The best-performing LLM requires only two input features: (1) bike lane length (km); and (2) the presence of schools, and achieves an average accuracy of 60%, just 4% below the OL model, and significantly better than the Naïve model and human survey respondents. Of interest, providing satellite images or additional quantitative input variables decreased the LLM performance. The results suggest LLMs can offer a scalable, data-efficient alternative for jurisdictions lacking extensive bicycle count data or modeling expertise.]]></description><pubDate>Tue, 27 Jan 2026 09:19:18 GMT</pubDate><guid>http://pubsindex.trb.org/view/2659296</guid></item><item><title>The Generative AI Challenge to Highway QA Integrity</title><link>http://pubsindex.trb.org/view/2657974</link><description><![CDATA[The rapid proliferation of accessible generative artificial intelligence (GenAI) presents a fundamental threat to the integrity of highway construction quality assurance (QA). Legacy systems are vulnerable to manipulation and fraud enabled by GenAI, especially through intuitive human–AI collaboration (“co-intelligence”). This paper introduces the Expert-AI Co-Research (EACR) framework—a structured way for a domain expert to work with GenAI across the research process—and applies it to conduct AI-augmented adversarial simulations that probe representative QA workflows. Four case studies provide, to the best of our knowledge, the first empirical demonstrations in the highway QA context that accessible GenAI can automate manipulation of testing data, including standardized binary file formats; help “game” specifications and procedures; generate or in-place edit photographic and technical QA imagery; and bypass built-in model safety features under adversarial prompting, with session-level “state persistence.” Harmful outputs were not operationalized, and operational details are withheld under responsible disclosure. Collectively, these findings show that GenAI democratized specialist-level manipulation, weakened the digital chain of custody, and elevated risk from isolated incidents to program-level decision risk. The paper contributes the EACR (framework) and AI-augmented adversarial simulations (method) as practical, repeatable tools for agencies to assess their systems, and provides the empirical basis for a necessary governance shift from “trust but verify” to “verify then trust”—emphasizing verifiable data origin, continuous monitoring, secure system design, and cross-disciplinary oversight. The approach offers agencies a concrete path to reassess and harden QA in the GenAI era.]]></description><pubDate>Mon, 26 Jan 2026 17:01:14 GMT</pubDate><guid>http://pubsindex.trb.org/view/2657974</guid></item><item><title>An Explainable Q-Learning Method for Longitudinal Control of Autonomous Vehicles</title><link>http://pubsindex.trb.org/view/2553289</link><description><![CDATA[Various artificial intelligence (AI) algorithms have been developed for autonomous vehicles (AVs) to support environmental perception, decision making and automated driving in real-world scenarios. Existing AI methods, such as deep learning and deep reinforcement learning, have been criticized due to their black box nature. Explainable AI technologies are important for assisting users in understanding vehicle behaviors to ensure that users trust, accept, and rely on AI devices. In this paper, an explainable Q-learning method for AV longitudinal control is proposed. First, AI control of AVs is realized by constructing a deep Q-network (DQN) with an intelligent driver model, with the control objective maximizing vehicle speed while preventing collisions. Then, a deep explainer for humans is developed via a Shapley additive explanation (SHAP), and a novel positive SHAP method that defines new base values is proposed to explain how individual state features contribute to decisions. Finally, statistical analyses and intuitive explanations are quantified based on SHAP tools to improve clarity. Elaborate numerical simulations are conducted to demonstrate the effectiveness of the proposed algorithm. The code is available at https://github.com/limeng-1234/Pos_Shap.]]></description><pubDate>Mon, 26 Jan 2026 14:17:10 GMT</pubDate><guid>http://pubsindex.trb.org/view/2553289</guid></item><item><title>A Three-Stage Decision-Making Method Based on Machine Learning for Preventive Maintenance of Airport Pavement</title><link>http://pubsindex.trb.org/view/2553284</link><description><![CDATA[The goal of preventative maintenance (PM) decision-making on airport pavements is to deploy the appropriate maintenance countermeasures at the correct time. This paper proposed a three-stage method for maintenance based on machine learning, which further refined the PM decision-making process. First, a pavement maintenance level model was developed using the PCA and PSO algorithm optimized SVM model. The model was then used to separate pavement maintenance into three categories: daily, PM, and major. Second, the DBSCAN and OPTICS were utilized to further divide the PM requirements finely. In order to implement the scientific decision-making of PM, suitable maintenance procedures were ultimately chosen based on the predominant damage kinds of the pavement units. The results showed that, when compared to the original SVM model, the classification accuracy of the PCA-PSO-SVM model was greatly improved, with total accuracy and accuracy of each class increasing by 10%, 41.7%, 4.6%, and 7.8%, respectively. When clustering the airport pavement performance dataset, OPTICS outperformed the DBSCAN technique. Four groups of PM demands were discovered by visualizing the best grouping levels after dimensionality reduction.]]></description><pubDate>Mon, 26 Jan 2026 14:17:10 GMT</pubDate><guid>http://pubsindex.trb.org/view/2553284</guid></item><item><title>Comparison of Machine Learning Approaches for Classifying and Evaluating Freeway Guide Signs</title><link>http://pubsindex.trb.org/view/2655744</link><description><![CDATA[Well-designed guide signs improve traffic safety by reducing reaction times and aiding decision-making. Current studies on guide sign assessments typically rely on simulated environments or expert subjective evaluations, which lack real data from actual driving conditions. Such data are crucial for accurately evaluating the practical effectiveness of guide signs from the driver’s perspective in complex and dynamic traffic environments. To systematically assess drivers’ physiological and psychological responses and identify guide sign deficiencies, this paper proposes an innovative approach by integrating emotional indicators, such as eye tracking, electroencephalography, and facial expressions, into guide sign effectiveness evaluation. These indicators help determine whether a guide sign is effective or ineffective. A comparison of several machine learning classification algorithms was undertaken to classify and evaluate freeway guide signs, using real-world road-testing data collected from Shijiazhuang, China. These indicators are used in the analysis, including accuracy, precision, recall, F1 score, and precision-recall curves. The extreme gradient boosting algorithm demonstrated the best performance, achieving 0.83 accuracy, 0.66 precision, 0.83 recall, 0.73 F1 score, and 0.71 area under the precision-recall curve for the minority class, while also excelling in the majority class. Percentage of eyelid closure, frowning, anger, and pupil area have particularly significant effects on the effectiveness of guide signs. Drivers preferred signs with place names, while inconsistent reliability and complex ramps led to lower ratings. This study provides empirical guidance for optimizing the effectiveness of freeway guide signs, emphasizing the critical role of physiological and psychological indicators in the evaluation process. It underscores that improving the clarity and reliability of guide signs is essential for enhancing road safety and promoting traffic efficiency.]]></description><pubDate>Fri, 23 Jan 2026 09:20:55 GMT</pubDate><guid>http://pubsindex.trb.org/view/2655744</guid></item><item><title>Do Riding Behavior, Speeding, and Law Adherence Differ between Professional and Nonprofessional Riders?</title><link>http://pubsindex.trb.org/view/2655733</link><description><![CDATA[Bangladesh has experienced a rapid rise in motorcycle use in recent years. The surge in nonprofessional and professional riders is as a result of demand, ride-sharing opportunities, self-employment, and affordability. Although riding behavior is identified as one of the most influential precursors of motorcycle safety, there is little research comparing the riding behaviors of professional and nonprofessional riders. This study, therefore, aims to differentiate professional and nonprofessional riders based on their distinct riding behavior. Data from 624 motorcycle riders were collected via online and face-to-face questionnaire surveys in Dhaka. Following the feature selection through mean decrease accuracy and mean decrease Gini, this study developed a random forest (RF) model to delve deeper into riders’ behavior. Furthermore, the SHapley Additive exPlanations-based feature was employed to determine the differential factors like overtaking vehicles from the wrong side, carrying passengers without helmets, no speed reduction at intersections, and riding without proper fitness. In addition, this study focused on finding differences in the two rider groups’ law adherence, alertness, and speed behaviors. Consequently, this enables policymakers to design data-driven targeted safety measures that more effectively address the unique risks the two rider groups pose. The findings suggest targeted educational interventions, awareness campaigns, and effective and strict enforcement of the Road Transportation Act of 2018 to improve safety practices. It also recommends purpose-built technology-based solutions, such as the use of wearable smart glasses and AI-integrated speed cameras, as well as engineering-based solutions, such as separating traffic moving at varying speeds and separate lanes for motorcycles.]]></description><pubDate>Thu, 22 Jan 2026 09:11:55 GMT</pubDate><guid>http://pubsindex.trb.org/view/2655733</guid></item><item><title>Investigating the Hybrid Approach of Anomaly Detection and EVT for Conflict-Based Crash Estimation Considering Variations in Safety Hierarchy</title><link>http://pubsindex.trb.org/view/2647077</link><description><![CDATA[The hybrid framework of extreme value theory (EVT) models and machine-learning-based anomaly detection methods has emerged as an effective approach for traffic-conflict-based crash estimation. This study investigates the generalizability of the hybrid framework across prevalent safety hierarchy models, namely pyramid-shaped and diamond-shaped hierarchies. Two anomaly detection methods—isolation forest (IForest) and minimum covariance determinant (MCD)—were integrated into the EVT modeling framework and compared with the conventional peak over threshold (POT) method. The methods were applied to two distinct datasets. One is the highD dataset corresponding to pyramid-shaped safety hierarchy, and the other is an intersection dataset corresponding to diamond-shaped hierarchy. The results reveal that IForest and MCD exhibit performance comparable to POT when applied to pyramid-shaped data with large sample sizes. However, these methods underperformed with diamond-shaped data because of its double-peak distribution, with performance improving when the data was reshaped to a pyramid shape. Furthermore, in the context of smaller datasets, an increase in the contamination rate for IForest and MCD was found to improve performance, surpassing POT in certain cases. Additionally, IForest displays sensitivity to the order of data, whereas MCD yields consistent outcomes regardless of the data arrangement.]]></description><pubDate>Thu, 08 Jan 2026 10:29:52 GMT</pubDate><guid>http://pubsindex.trb.org/view/2647077</guid></item><item><title>Recognition of Truck Brands Using Fused Deep Neural Networks Based on Transfer Learning</title><link>http://pubsindex.trb.org/view/2647075</link><description><![CDATA[Long-distance truck transportation often results in stained, faded, or blurry frontal images, leading to fluctuations in image resolution and feature degradation. To enable high-precision truck brand recognition in challenging highway scenarios, this study evaluates convolutional neural networks (CNNs) enhanced by transfer learning (TL). First, four typical neural networks, such as InceptionV3 based on TL, Xception based on TL, Xception based on TL and DenseNet201 based on TL, are exploited for the recognition of truck brands. Second, a new network architecture is proposed, a fused deep neural network (FDNN), based on transfer learning for the recognition of truck brands (FDNN–TL–RTB), which integrates the convolution features of the last layer of InceptionV3–TL–RTB, Xception–TL–RTB, and DenseNet-201–TL–RTB networks based on the tandem fusion rules. Finally, the comparative experiments on CNNs are carried out using the Truck Brands Data Sets of Southeast University. The data set was obtained using continuous capture with cameras installed on the highway to obtain data under various conditions. The experimental results demonstrate that the proposed FDNN–TL–RTB network achieves a superior recognition accuracy of 98.16% on the test set. Although Shidai vehicles presented the most significant classification challenges among the 23 truck brand categories, the FDNN–TL–RTB method achieved a 95% recognition accuracy for this category. The high accuracy and robustness of the model highlight its significant potential for practical applications in intelligent transportation systems, such as automated toll collection and traffic flow monitoring.]]></description><pubDate>Thu, 08 Jan 2026 10:29:52 GMT</pubDate><guid>http://pubsindex.trb.org/view/2647075</guid></item><item><title>Retrieval Augmented Generation-Based Large Language Models for Bridging Transportation Cybersecurity Legal Knowledge Gaps</title><link>http://pubsindex.trb.org/view/2647045</link><description><![CDATA[With the advancement of connected and automated transportation systems, a growing need has emerged for regulatory authorities to update existing laws and statutes, as well as create new ones, to address future implications of connectivity and automation. Specifically, ensuring proper engagement with cybersecurity and data privacy challenges for connected and automated transportation systems will require a comprehensive legal framework at both the federal and state levels. To help policymakers achieve this, a retrieval augmented generation (RAG)-based large language model (LLM) framework, Transportation Cybersecurity and Resiliency (TraCR) AI, has been developed in this study, focusing on extracting pertinent information from existing legislation based on inquiries and crafting LLM-generated responses to highlight potential loopholes for further scrutiny. This study primarily aims to mitigate the hallucinations caused in the domain of LLMs by developing a curated knowledge base of legislative documents and an associated question–answer dataset to improve the effectiveness of query results. This RAG-based framework extracts relevant information to improve the specificity of answers and aids the LLM in providing factually accurate responses with increased reliability. Our analyses reveal that the presented RAG-based framework can aid legislative analysis by generating queries for particular questions and responses. We also compare our RAG framework-generated responses with commercially available LLMs to demonstrate the effectiveness of our approach. TraCR AI outperforms leading commercial LLMs across four distinct metrics, that is, AlignScore, ParaScore, BERTScore, and ROUGE score. This highlights that integrating RAG allows LLMs to produce more factually accurate and up-to-date responses than standalone LLMs. This approach to domain-specific LLM development will improve the quality of legislative analysis that can be used to aid policymakers in meeting the challenge of updating legal codes in accordance with emerging technologies.]]></description><pubDate>Wed, 07 Jan 2026 14:58:17 GMT</pubDate><guid>http://pubsindex.trb.org/view/2647045</guid></item><item><title>Demand Forecasting of Shared Bicycles at Subway Entrances using SSA-LSTM-RF</title><link>http://pubsindex.trb.org/view/2647080</link><description><![CDATA[Shared bicycles are vital for solving the “last-mile” problem in green and sustainable urban development. However, inaccurate demand forecasting causes severe supply-demand imbalances, especially around subway stations. To address this, we propose an improved sparrow search algorithm-long short-term memory neural network-random forest (SSA-LSTM-RF) hybrid model for accurate shared-bicycle demand prediction at subway entrances. The proposed predictive model is founded on the LSTM neural network and RF. SSA is effectively employed to optimize the hyperparameters of these models, and, finally, the least squares weighting method is utilized to integrate the two models effectively. Taking the shared-bicycle order data from Zhu Guang Station in Nanshan District, Shenzhen, China, as a case study, the SSA-LSTM-RF model was benchmarked against traditional seasonal autoregressive integrated moving average and light gradient boosting machine models. The experimental results revealed that the SSA-LSTM-RF model exhibited the lowest mean absolute error and root mean square error, which were 9.8783 and 14.0117 respectively, and the highest coefficient of determination of 0.9943. These results demonstrate the superiority of the SSA-LSTM-RF model in predicting shared-bicycle demand. The model was further applied to predict the shared-bicycle demand at all subway entrances within the district. Results indicated that the model achieved a prediction accuracy of over 99% for high-demand stations, and that it performed well on datasets with high demand and significant temporal variations. This research offers theoretical support for shared-bicycle system planning, improving subway services and the urban transportation ecosystem.]]></description><pubDate>Wed, 07 Jan 2026 14:58:17 GMT</pubDate><guid>http://pubsindex.trb.org/view/2647080</guid></item><item><title>Comparative Evaluation of Machine Learning Models for Pavement Roughness Prediction</title><link>http://pubsindex.trb.org/view/2646127</link><description><![CDATA[This study aims to develop and compare different models for predicting the International Roughness Index (IRI) of 419 long-term pavement performance pavement sections from seven states in the United States. The IRI is a crucial metric used to evaluate pavement roughness, ranging from very smooth to very rough surfaces. Machine learning models have gained popularity in predicting IRI as a result of their ability to analyze large volumes of data, improving accuracy and providing cost-effective solutions for pavement management and maintenance. The developed models are generalized linear model (GLM), support vector machine regression, multivariate adaptive regression splines, artificial neural network (ANN), and extreme gradient boosting (XGBoost). The XGBoost model outperformed the other models with the lowest root mean square error and the highest R². The GLM model showed good performance for lower values of the roughness index while it underpredicted higher values. After regularization and feature selection, the input variables that were common to all models included age, structural number, deflection at 1.5 times of pavement thickness, traffic load, precipitation, and maintenance and rehabilitation history. By utilizing these models, pavement engineers can make informed decisions, allocate resources efficiently, and prioritize maintenance activities based on accurate predictions of pavement roughness.]]></description><pubDate>Tue, 30 Dec 2025 13:51:27 GMT</pubDate><guid>http://pubsindex.trb.org/view/2646127</guid></item></channel></rss>