International Journal of Transportation Science and Technology最新文献

{"title":"Data-driven analysis of run-level bus alighting patterns for accurate predictions and operational efficiency","authors":"Payel Roy,&nbsp;Karthik K. Srinivasan","doi":"10.1016/j.ijtst.2025.06.006","DOIUrl":"10.1016/j.ijtst.2025.06.006","url":null,"abstract":"<div><div>Accurate prediction of bus passengers alighting at each stop during each run is crucial to improve ridership and implement destination-specific demand-responsive solutions like last-mile connectivity (LMC) and multimodal transfers. However, existing studies often disregard empirical distribution, heterogeneity, and heteroscedasticity of alighting patterns at disaggregate run-level resolution. Machine learning (ML) models, although capable of modeling complex relationships, are often replaced by traditional methods due to interpretability needs in policymaking. Therefore, this paper investigates factors influencing bus alighting behavior using ML techniques, namely decision tree, random forest, and gradient boosting (GB), and compares against a traditional ordinary least squares (OLS) model to evaluate prediction accuracy, variable impacts, and robustness at the run level with electronic ticketing data from Chennai, India. Results show that GB outperforms OLS with a 51% improvement in mean squared error (MSE) (from 2.50 to 1.23) and a 41% improvement in root mean squared error (RMSE) (from 3.76 to 2.21), respectively. It also effectively addresses zero inflation, heterogeneity, and heteroscedasticity without data transformation. Passenger occupancy, a real-time operational variable, explained 29% of total variability, capturing dynamic conditions that static variables cannot. The ‘neighborhood attraction points’ category accounts for 6.5% of the total variation, revealing trip-chaining and substitution effects at not only the current but also preceding and subsequent stages. Coarser resolution data (e.g., hourly/daily) cause inaccurate generalizations and reduce variables’ significance, emphasizing the need for finer run-level resolution for robust predictions. A practical application of the proposed model demonstrates that increasing ordinary bus services from the current 7–9% optimizes ridership by balancing enhanced users’ affordability with minimized revenue loss for transit operators.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"23 ","pages":"Pages 115-136"},"PeriodicalIF":5.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149078751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Selecting and optimizing distress feature descriptors from multi-source images for pavement distress classification","authors":"Aidi Wang,&nbsp;Hong Lang,&nbsp;Yichuan Peng,&nbsp;Xinting Chen,&nbsp;Jinsong Qian","doi":"10.1016/j.ijtst.2025.05.003","DOIUrl":"10.1016/j.ijtst.2025.05.003","url":null,"abstract":"<div><div>Accurate and efficient early detection of pavement distress is crucial, as it prevents further deterioration, reduces repair costs, and enhances driving safety by mitigating hazards like potholes and cracks. However, automatically detecting pavement distress from image samples remains challenging due to complex asphalt textures, including variations in aggregate size, surface wear, and non-distress-related artifacts like shadows and tire marks. This study addresses these challenges by screening and optimizing texture descriptors to characterize local structures and patterns of image regions containing pavement distress. An automated detection system is developed using four categories of texture descriptors: (1) original image-based parameters, (2) multi-resolution analysis-based texture features, (3) local binary pattern (LBP) and their variants-based texture features, and (4) gray-level co-occurrence matrix (GLCM)-based texture features. Unlike previous studies that rely on a single type of descriptor, this integrated approach processes both two-dimensional (2D) intensity images and three-dimensional (3D) pavement surface models, capturing global and local texture patterns to improve robustness and feature discriminability for distress classification. A total of 196 optimal feature descriptors are selected to represent images from both imaging modalities, forming the basis of a distress classification model. A dataset comprising 4 932 samples across 8 distress categories, with moderate class imbalance, is constructed to compare the performance of random forest (RF), gradient boosting decision tree (GBDT), multilayer perceptron (MLP), and support vector machine (SVM) in distress classification. Class-weighted training and balanced performance metrics were used to mitigate class imbalance effects and ensure robust model evaluation. Among these, the SVM classifier demonstrates the best predictive performance and robustness. Extensive experiments, including feature combination analysis and ablation studies, identify the optimal feature set comprising original image-based, multi-resolution analysis-based, and LBP variants-based descriptors. This feature set achieves a classification accuracy rate (CAR) of 0.896, an F1-score of 0.894, and an area under receiver operating characteristic curve (AUC) of 0.938 across all distress categories. The selected descriptors effectively differentiate pavement distress types, providing a reliable automated classification system.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 383-409"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Stochastic multi-objective optimization for dynamic timetable and track allocation at high-speed railway hubs","authors":"Bing Yan,&nbsp;Lu Huang,&nbsp;Chen Wan,&nbsp;Siyuan Qu,&nbsp;Xiaodong Fan,&nbsp;Xiaolei Zou","doi":"10.1016/j.ijtst.2025.02.011","DOIUrl":"10.1016/j.ijtst.2025.02.011","url":null,"abstract":"<div><div>Train scheduling and track allocation are crucial for minimizing passenger flow conflicts, ensuring safety, and enhancing travel experience at railway hubs. This study presents a collaborative optimization model for railway hub operations, focusing on improving passenger transfer efficiency and minimizing arrival-departure track (ADT) utilization costs. A multi-objective hybrid model, referred to as COTADT, is developed to address spatiotemporal and stochastic constraints while balancing these two objectives. To solve this model efficiently, a customized augmented epsilon-constraint (CAEC) algorithm is introduced, utilizing augmented constraints and interval partitioning to generate Pareto-optimal solutions. The approach is validated with real-world data from the Hangzhou Hub, yielding significant improvements, including a 23.149% reduction in passenger transfer costs and a 7.101% reduction in ADT utilization costs. Comparative experiments demonstrate that CAEC outperforms heuristic algorithms in both solution quality and computational efficiency. This research provides a robust, scalable framework for enhancing operational performance and passenger experience at railway hubs.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 1-17"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545070","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"HCT-LSTM: integrating habit, current situation, and tendency for driving safety state prediction in adverse weather conditions","authors":"Benmin Liu,&nbsp;Chengxin Li,&nbsp;Bingyu Shi,&nbsp;Runzhu Deng","doi":"10.1016/j.ijtst.2025.04.005","DOIUrl":"10.1016/j.ijtst.2025.04.005","url":null,"abstract":"<div><div>A model known as habit + current situation + tendency (HCT) has been developed for evaluating driving conduct, with the consideration of the diversity among drivers and the persistent characteristics of driving behaviors under inclement weather conditions. The long short-term memory (LSTM) networks were utilized for forecasting driving safety levels. Initially, data were gathered from driving simulation trials, and an HCT indicator system was formulated based on driving patterns across different scenarios, encompassing 12 habitual descriptive parameters, 4 current state descriptive parameters, and 4 fluctuation descriptive parameters. Subsequent application of <em>K</em>-means clustering led to the classification of driving safety into three levels: dangerous, moderately dangerous, and safe, while driving stability was segmented into six distinct categories. Principal component analysis (PCA) was then utilized to streamline the habit-related parameters. Ultimately, an LSTM model was implemented to construct a predictive framework for safety levels, enabling real-time risk management for highway driving in adverse weather conditions. The model exhibited an overall accuracy and a recall rate of 88.59% and 94.24%, respectively, establishing its efficacy as a computational tool for immediate risk mitigation.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 247-262"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intelligent vehicle routing for stochastic service times: a grouping evolution strategy approach","authors":"Sepinoud Rezvanian,&nbsp;Ali Husseinzadeh Kashan,&nbsp;Alireza Rezvanian,&nbsp;Alireza Sabzevari","doi":"10.1016/j.ijtst.2025.04.009","DOIUrl":"10.1016/j.ijtst.2025.04.009","url":null,"abstract":"<div><div>After-sales service companies often need to transport and distribute their products, as well as provide services like installation and setup. In this paper, we study a vehicle routing problem (VRP) that considers the possibility of on-site servicing by installers, with stochastic service times. We formulate the distribution of products to customers as a VRP and extract the route plan for installers by solving a traveling salesman-type problem. To model the stochastic service times, we use a chance constraint while minimizing the total cost for both vehicles and installers. We develop a grouping evolution strategy (GES) algorithm equipped with various heuristics to generate efficient solutions for large-scale problems. Numerical results are reported from a case study using existing data for the problem. The computational experiments demonstrate that GES can produce near-optimal solutions for operational use within a reasonable amount of time. In summary, our study shows the effectiveness of GES in solving the VRP with on-site stochastic service times, which is relevant to many after-sales service companies.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 315-334"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Socially responsible pay for priority traffic control system for e-commerce delivery vehicles","authors":"Slavica Gavric,&nbsp;Ismet Goksad Erdagi,&nbsp;Aleksandar Stevanovic","doi":"10.1016/j.ijtst.2025.04.008","DOIUrl":"10.1016/j.ijtst.2025.04.008","url":null,"abstract":"<div><div>This research proposes a socially responsible pay-for-priority traffic control system for e-commerce delivery vehicles (EDVs). The system aims to maximize revenue collection while preserving existing traffic conditions through two key components: 1) an annual base fee paid by EDVs for road usage, and 2) a fuzzy logic-based pay-for-priority system allowing EDVs to request signal priority for a nominal fee in specific cases. The system was implemented and tested using microsimulation on an 11-intersection corridor in Chattanooga, Tennessee. The results demonstrate that the system reduces EDVs’ delays by 5%–10% and travel times by 4%–5%, while maintaining constant delays for passenger cars across various traffic demand and EDV penetration scenarios. A cost-benefit analysis of the proposed system indicates that the system would be economically viable for all scenarios, with the annual base fee generating the majority of revenue. The study also presents a novel approach to modeling vehicles with multiple random deliveries in microsimulation. This research contributes to addressing the externalities of increased e-commerce deliveries by proposing an innovative charging mechanism that can generate revenue for infrastructure maintenance while improving EDV efficiency. Future work should examine the system’s performance in grid networks and consider scenarios where deliveries cause lane closures.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 297-314"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545091","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing autonomous vehicle safety: an integrated ensemble learning-logit model for accident severity prediction and analysis","authors":"Zhenyu Zhao,&nbsp;Tingyu Liu,&nbsp;Lanfang Zhang,&nbsp;Shuke Xie,&nbsp;Hao Jin","doi":"10.1016/j.ijtst.2025.03.005","DOIUrl":"10.1016/j.ijtst.2025.03.005","url":null,"abstract":"<div><div>As automated driving technology advances rapidly, ensuring its safety has become a crucial area of research and practice. This study presents an integrated ensemble learning-logit model (IELLM) designed to enhance the precision in predicting the severity of accidents involving automated vehicles and to explain the factors influencing these accidents. Initially, the model utilizes ensemble learning techniques, integrating multiple machine learning algorithms to predict accident severity and identify key contributing factors accurately. Then, an ordered logit regression model is1 employed to examine critical variables affecting accident severity, such as the level of the automated driving system, pre-collision vehicle speed, and collision site. The findings indicate that vehicles with higher levels of automation (SAE Level 3–5) perform significantly better in accidents than those with lower automation, highlighting the potential of advanced automation technologies to enhance road safety. Moreover, the research identifies high speeds and specific collision types as significant factors contributing to increased accident severity. Through a systematic analysis of these variables, the study deepens the understanding of the characteristics of accidents involving automated vehicles and provides a scientific basis for formulating relevant safety policies and vehicle designs. Based on these findings, recommendations are proposed to improve the safety of automated vehicles, including further research on advanced automated systems, enhanced professional training for drivers, optimized vehicle structural design, and the implementation of stricter speed control measures. These initiatives aim to reduce the incidence and severity of accidents involving automated vehicles, contributing to a safer road traffic environment.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 111-125"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Stochastic optimization model for port infrastructure planning considering uncertainties in wave and vessel arrival times","authors":"Rajin Sihombing,&nbsp;Ricky Lukman Tawekal,&nbsp;Muslim Muin,&nbsp;Febri Zukhruf","doi":"10.1016/j.ijtst.2025.03.004","DOIUrl":"10.1016/j.ijtst.2025.03.004","url":null,"abstract":"<div><div>This paper proposes a stochastic optimization model that considers not only the uncertainties in wave and vessel arrival times but also sedimentation and tidal occurrence. Those factors impact the vessel’s operational cost, influencing the seaport’s competitiveness. Therefore, seaport operators need to decide the reference vessel size, dredging maintenance actions, and infrastructure required to ensure the smooth operation of the vessel. The optimization model is solved by invoking a variant of the genetic algorithm (GA). Case studies based on actual seaports in Indonesia are conducted to illustrate the model’s application and results clearly. It was found that the uncertainties of wave and vessel arrival are responsible for deviations in expected operational costs, while increasing infrastructure costs minimizes these deviations and reduces the expected operational costs.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 94-110"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"COM-Roundabout: the first smart commutable and self-regulating roundabout for HDVs and CAVs","authors":"Marco Guerrieri,&nbsp;Masoud Khanmohamadi","doi":"10.1016/j.ijtst.2025.04.003","DOIUrl":"10.1016/j.ijtst.2025.04.003","url":null,"abstract":"<div><div>In the last decades, many new types of roundabouts and traffic regulation systems have been introduced in several countries. However, some issues related to their performance and safety remain. This article presents a new concept of self-regulating smart roundabout recently ideated: the commutable roundabout (COM-Roundabout), where the number of active entry and circulating lanes is dynamically adjusted in real time, depending on the measured traffic demand intensity. The COM-Roundabout utilizes vehicle and queue detectors, smart cameras, light-emitting diode (LED) road markers, and variable message signs (VMSs) to guide traffic into appropriate lanes dynamically. Performance in terms of capacity, delay, level of service (LOS) and predicted accident rates is evaluated in many traffic scenarios, based on an extensive dataset of one million origin-destination (O/D) traffic matrices (<em>M</em>_O/D), considering both human-driven vehicles (HDVs) and cooperative autonomous vehicles (CAVs), applying robust and widely used estimation methods (HCM and HSM models). This research demonstrates that the COM-Roundabout could be used to optimize real-time performance (e.g., capacity and LOS) and safety thanks to the numerous layouts and traffic control programmes adaptable to specific traffic conditions.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 213-230"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep charge-fusion model: advanced hybrid modelling for predicting electric vehicle charging patterns with socio-demographic considerations","authors":"Muhammed Cavus,&nbsp;Huseyin Ayan,&nbsp;Margaret Bell,&nbsp;Oluwole K. Oyebamiji,&nbsp;Dilum Dissanayake","doi":"10.1016/j.ijtst.2025.03.002","DOIUrl":"10.1016/j.ijtst.2025.03.002","url":null,"abstract":"<div><div>This study examines electric vehicle (EV) charging behaviors across 5 898 participants from various socio-demographic backgrounds, using both statistical analysis and deep learning models. Exploratory data analysis (EDA) and analysis of variance (ANOVA) show that factors such as gender, education, employment status, household size, income, and housing type significantly influence charging behaviors at home, work, and fast-charging stations. Education notably impacts home charging (<em>F</em> = 2.54, <em>p</em> = 0.04), while household income strongly affects fast charging behavior (<em>F</em> = 5.34, <em>p</em> = 0.001). A hybrid deep learning model, deep charge-fusion, was developed by combining convolutional neural networks (CNNs), long short-term memory (LSTM) networks, and extreme gradient boosting (XGBoost). CNNs capture spatial patterns for regional charging, while LSTMs model temporal dependencies in charging patterns. XGBoost adds efficiency by managing structured data and preventing overfitting through regularization. Dropout layers were introduced in the LSTM network to reduce overfitting. The deep charge-fusion model achieved an <em>R</em>² value of 0.81 for predicting fast-charging behaviors, outperforming standalone models such as gated recurrent unit (GRU)-based (<em>R</em>² = 0.51) and LSTM-based (<em>R</em>² = 0.43) models. It also achieved an <em>R</em>² of 0.81 for home charging predictions. These results underscore the role of socio-demographic factors and demonstrate that hybrid models significantly improve predictive accuracy, with implications for EV infrastructure planning and energy management strategies.</div></div>","PeriodicalId":52282,"journal":{"name":"International Journal of Transportation Science and Technology","volume":"22 ","pages":"Pages 37-60"},"PeriodicalIF":5.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148545057","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}