Parseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Driver Lane-Change Prediction on Freeways Using Long Short-Term Memory Network and Support Vector Machines Based on Microscopic DataDriver Lane-Change Prediction on Freeways Using Long Short-Term Memory Network and Support Vector Machines Based on Microscopic Data5635565123680210.22119/jte.2025.534186.2742FAMohammadManiatDepartment of Civil Engineering, SR.C., Islamic Azad University, Tehran, Iran0000-0002-0515-4821MahmoudSaffarzadehProfessor,Department of Civil and Environmental Engineering, Tarbiat Modares University, Tehran, Iran0000-0002-9713-2776HasanZoghiAssistant Professor, Department of Civil Engineering, Islamic Azad University, Karaj, IranسرکارSarkarAssistant Professor, Department of Civil Engineering, SR.C., Islamic Azad University, Tehran, Iran0000-0003-4449-544XJournal Article20250713Lane changing is one of the most critical driving maneuvers, playing a significant role in both traffic safety and flow efficiency. The aim of this study is to develop and compare machine learning models for short-term prediction of drivers’ lane-change behavior on freeways. For this purpose, high-resolution microscopic data from the Next Generation Simulation (NGSIM) dataset were utilized, and two predictive models—Long Short-Term Memory (LSTM) neural networks and Support Vector Machines (SVM), were designed and evaluated. Predictions were performed for time horizons ranging from 1 to 5 seconds prior to the lane-change maneuver in order to assess the effect of prediction horizon on model performance. Evaluation metrics, including Mean Squared Error (MSE), Recall, and F1-score, were applied to assess model accuracy. The results indicate that the LSTM model significantly outperformed SVM across all evaluation criteria and achieved the highest accuracy with the lowest error in short-term horizons (1–5 seconds). Furthermore, statistical analysis revealed that lane-change maneuvers were more frequent in the right lanes of the freeway. The findings of this study can contribute to the design of advanced driver-assistance systems (ADAS) and the development of accurate predictive algorithms for semi-autonomous and autonomous vehicles.Lane changing is one of the most critical driving maneuvers, playing a significant role in both traffic safety and flow efficiency. The aim of this study is to develop and compare machine learning models for short-term prediction of drivers’ lane-change behavior on freeways. For this purpose, high-resolution microscopic data from the Next Generation Simulation (NGSIM) dataset were utilized, and two predictive models—Long Short-Term Memory (LSTM) neural networks and Support Vector Machines (SVM), were designed and evaluated. Predictions were performed for time horizons ranging from 1 to 5 seconds prior to the lane-change maneuver in order to assess the effect of prediction horizon on model performance. Evaluation metrics, including Mean Squared Error (MSE), Recall, and F1-score, were applied to assess model accuracy. The results indicate that the LSTM model significantly outperformed SVM across all evaluation criteria and achieved the highest accuracy with the lowest error in short-term horizons (1–5 seconds). Furthermore, statistical analysis revealed that lane-change maneuvers were more frequent in the right lanes of the freeway. The findings of this study can contribute to the design of advanced driver-assistance systems (ADAS) and the development of accurate predictive algorithms for semi-autonomous and autonomous vehicles.https://jte.sinaweb.net/article_236802_aaa1b825df9ce04f2b502c4896f66518.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Examining Behavioral Factors Affecting the Frequency of Motorcycle Accidents and Near-Crash Events (Case Study: Isfahan)Examining Behavioral Factors Affecting the Frequency of Motorcycle Accidents and Near-Crash Events (Case Study: Isfahan)5653567424412910.22119/jte.2025.535016.2743FAMehdiFallah TaftiAssociate Professor, Civil Engineering Department, Yazd University, Yazd, Iran0000-0001-6844-3316MojtabaRajabi-BahaabadiAssistant Professor, Civil Engineering Department, Yazd University, Yazd, Iran0000-0002-1392-5729َAhmadMohajeriPh.D. Student, Civil Engineering Department, Yazd University, Yazd, IranMohammad AminJaziM.Sc., Civil Engineering Department, Yazd University, Yazd, IranJournal Article20250730Understanding the behavioral patterns that influence motorcyclist accidents plays a significant role in developing traffic safety programs. The objective of this research was to investigate the relationship between motorcyclists' behavioral factors and the rate of accidents and near-miss incidents among riders in Isfahan. The standard "Motorcycle Rider Behavior" questionnaire was revised, with several items added to suit Iranian conditions. After data collection, Exploratory Factor Analysis identified four main factors: "Risky, Showy, and Impulsive Behaviors", "Intentional Violations", "Non-adherence to Safety Equipment Principles," and "Perceptual and Cognitive Errors". To evaluate the impact of these factors, two count regression models namely, "number of accidents in the last two years" and "number of near-miss incidents in the last three months," were developed. The results indicated that in both models, all behavioral factors were significant. However, the "Intentional Violations" and "Risky, Showy, and Impulsive Behaviors" factors had the greatest impact, with standardized coefficients of 0.8 and 0.6 in the first model, and 0.49 and 0.44 in the second. Among the 24 behaviors examined, "Failure to use turn signals" and "Changing lanes without checking mirrors" (with an average score of 3.3 out of 5) were identified as the most common high-risk behaviors. Behaviors such as "Using a mobile phone," "Failure to observe the speed limit," "Overtaking from the right," and "Weaving between lanes" (with an average score of 2.9 out of 5) ranked next. The findings of this research can be used to develop educational interventions, oversight measures, and policy-making for enhancing the safety of motorcyclists.Understanding the behavioral patterns that influence motorcyclist accidents plays a significant role in developing traffic safety programs. The objective of this research was to investigate the relationship between motorcyclists' behavioral factors and the rate of accidents and near-miss incidents among riders in Isfahan. The standard "Motorcycle Rider Behavior" questionnaire was revised, with several items added to suit Iranian conditions. After data collection, Exploratory Factor Analysis identified four main factors: "Risky, Showy, and Impulsive Behaviors", "Intentional Violations", "Non-adherence to Safety Equipment Principles," and "Perceptual and Cognitive Errors". To evaluate the impact of these factors, two count regression models namely, "number of accidents in the last two years" and "number of near-miss incidents in the last three months," were developed. The results indicated that in both models, all behavioral factors were significant. However, the "Intentional Violations" and "Risky, Showy, and Impulsive Behaviors" factors had the greatest impact, with standardized coefficients of 0.8 and 0.6 in the first model, and 0.49 and 0.44 in the second. Among the 24 behaviors examined, "Failure to use turn signals" and "Changing lanes without checking mirrors" (with an average score of 3.3 out of 5) were identified as the most common high-risk behaviors. Behaviors such as "Using a mobile phone," "Failure to observe the speed limit," "Overtaking from the right," and "Weaving between lanes" (with an average score of 2.9 out of 5) ranked next. The findings of this research can be used to develop educational interventions, oversight measures, and policy-making for enhancing the safety of motorcyclists.https://jte.sinaweb.net/article_244129_ca7e21a36d12fff53ec69d904179aca4.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Developing Graph Centrality Concepts in Analyzing and Prioritizing Transportation Network ComponentsDeveloping Graph Centrality Concepts in Analyzing and Prioritizing Transportation Network Components5675569824551410.22119/jte.2026.536860.2744FAAmirHosseinpanahiPhD student in civil engineering, transportation planning, Islamic Azad University, central Tehran branch, Tehran, IranAlirezaMahpourTransportation Planning, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University, Tehran0000-0001-9454-5906HoomanRahimiAssistant Professor of Transportation Planning, Department of civil engineering ShQ.C,Islamic azad university,Shahr-e-Qods,IranHassanDivandariAssistant Professor of Transportation Planning, Department of civil engineering ShQ.C,Islamic azad university,Shahr-e-Qods,IranJournal Article20250725Identifying key components of urban transportation networks to enhance resilience, reduce vulnerability, and optimize traffic flow is a fundamental issue in transportation planning. In this study, an analytical framework based on graph-theoretical centrality concepts and matrix–tensor structures was developed to determine and prioritize the relative importance of network components. To this end, three incidence matrices—route–street, street–intersection, and route–intersection—were defined, and a three-dimensional tensor was constructed to represent the multilayer relationships of the network simultaneously. Degree, betweenness, and eigenvector centrality measures were then computed based on these structures and integrated through normalization and weighted aggregation. The proposed framework was implemented on the benchmark Nguyen–Dupuis network consisting of 13 nodes and 38 links. The results indicated that nodes 6, 11, and 5 exhibit the highest composite centrality values, and selecting the top five nodes covers 92.3% of the network routes. Sensitivity analysis further revealed that the tensor-based criterion can highlight inter-layer nodes that are not identified by classical methods. The findings suggest that the combined use of incidence matrices and tensor structures provides deeper insight into the functional organization of transportation networks and can support more effective strategies for maintenance prioritization, alternative route design, and system resilience enhancement.Identifying key components of urban transportation networks to enhance resilience, reduce vulnerability, and optimize traffic flow is a fundamental issue in transportation planning. In this study, an analytical framework based on graph-theoretical centrality concepts and matrix–tensor structures was developed to determine and prioritize the relative importance of network components. To this end, three incidence matrices—route–street, street–intersection, and route–intersection—were defined, and a three-dimensional tensor was constructed to represent the multilayer relationships of the network simultaneously. Degree, betweenness, and eigenvector centrality measures were then computed based on these structures and integrated through normalization and weighted aggregation. The proposed framework was implemented on the benchmark Nguyen–Dupuis network consisting of 13 nodes and 38 links. The results indicated that nodes 6, 11, and 5 exhibit the highest composite centrality values, and selecting the top five nodes covers 92.3% of the network routes. Sensitivity analysis further revealed that the tensor-based criterion can highlight inter-layer nodes that are not identified by classical methods. The findings suggest that the combined use of incidence matrices and tensor structures provides deeper insight into the functional organization of transportation networks and can support more effective strategies for maintenance prioritization, alternative route design, and system resilience enhancement.https://jte.sinaweb.net/article_245514_4ef1830e8ed513b6733808be56492598.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Identifying and Segmentation Key Customers Based on Value Creation Criteria in Industrial Markets (Case Study: International Road Transport Companies)Identifying and Segmentation Key Customers Based on Value Creation Criteria in Industrial Markets (Case Study: International Road Transport Companies)5699572624413010.22119/jte.2026.550543.2747FANaserSheikhDepartment of Business Administration, NT.C., Islamic Azad University, Tehran, IranHamidrezaYazdaniDepartment of Leadership and Human Capital, College of Management, University of Tehran, Tehran, IranAbdol-HosseinKarampourDepartment of Business Administration, NT.C., Islamic Azad University, Tehran, IranJournal Article20251002The international road transport industry, as one of the core pillars of the global supply chain, is facing challenges in terms of competitiveness and customer relationship management. In this context, identifying and clustering key customers based on value creation criteria represents an innovative approach to enhancing efficiency and strengthening companies’ competitive advantage. However, the multi-layered dimensions of customer value creation have received limited attention from researchers. Accordingly, the purpose of this study was to identify and cluster key customers in industrial markets based on value creation criteria. This research is considered a mixed-method study. In the qualitative phase, customer value creation criteria were elicited, while in the quantitative phase, the customer lifetime value (CLV) of key industrial customers was first calculated using a formulated equation, and subsequently, clustering was performed based on the elicited criteria. In the qualitative section, the sample consisted of 14 experts selected through purposive non-probability sampling. In the quantitative section, purposive and convenience sampling was employed to calculate the CLV of key customers, involving 42 companies. The K-Means method was then applied for clustering. The findings revealed that key industrial customers were grouped into three clusters: (1) Golden Low-Interaction Customers—customers with high revenue per transaction but fewer transactions compared to other clusters; (2) Silver Growth-Potential Customers—customers with a high number of transactions and continuous engagement, generating moderate revenue yet demonstrating a high renewal rate; and (3) Bronze High-Interaction Customers—customers with the highest number of transactions but very low revenue.The international road transport industry, as one of the core pillars of the global supply chain, is facing challenges in terms of competitiveness and customer relationship management. In this context, identifying and clustering key customers based on value creation criteria represents an innovative approach to enhancing efficiency and strengthening companies’ competitive advantage. However, the multi-layered dimensions of customer value creation have received limited attention from researchers. Accordingly, the purpose of this study was to identify and cluster key customers in industrial markets based on value creation criteria. This research is considered a mixed-method study. In the qualitative phase, customer value creation criteria were elicited, while in the quantitative phase, the customer lifetime value (CLV) of key industrial customers was first calculated using a formulated equation, and subsequently, clustering was performed based on the elicited criteria. In the qualitative section, the sample consisted of 14 experts selected through purposive non-probability sampling. In the quantitative section, purposive and convenience sampling was employed to calculate the CLV of key customers, involving 42 companies. The K-Means method was then applied for clustering. The findings revealed that key industrial customers were grouped into three clusters: (1) Golden Low-Interaction Customers—customers with high revenue per transaction but fewer transactions compared to other clusters; (2) Silver Growth-Potential Customers—customers with a high number of transactions and continuous engagement, generating moderate revenue yet demonstrating a high renewal rate; and (3) Bronze High-Interaction Customers—customers with the highest number of transactions but very low revenue.https://jte.sinaweb.net/article_244130_b9ade1dadf155b75c9b7cf62777a6898.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Investigating the Impact of Music on Driver Safety with a Focus on Biometric Indicators and Personality TraitsInvestigating the Impact of Music on Driver Safety with a Focus on Biometric Indicators and Personality Traits5727574023680310.22119/jte.2025.553437.2750FAAlirezaSadeghiM.Sc. in Transportation Planning, Department of Civil and Environmental Engineering, Tarbiat Modares University, Tehran, IranSaeidSherafatipourAssistant Professor, Department of Transportation Planning, Faculty of Civil and Environmental Engineering, Tarbiat Modares University, Tehran, Iran0009-0004-2291-1718HojatollahFarahaniAssociate Professor, Department of Psychology, Tarbiat Modares University, Tehran, Iran0000-0002-9799-7008Journal Article20251015This research investigates driving performance and physiological arousal as influenced by structural features of music using a within-subject design with 50 licensed participants. Fifteen music trials, created using fractional factorial design with factors including manipulated musical emotion, musical tempo, and volume, are conducted in the BeamNG-Drive simulator. The physiological measure SR was logged and scaled to create a common Arousal Index. The driving variables logged include driving errors, speeding pattern, lateral deviation from lanes, and reaction time<span dir="RTL" lang="FA">.</span> PCA on music-attitude data validated data adequacy, while k-means clustering extracted five distinct types of attitudes on music. Cluster analysis on features related to performance and physiology also extracted four stable patterns of responses from “low-error/low-arousal” to “high-error/high-arousal.” On within-subject statistical tests, musical condition had significant effects on both error rates and levels of arousal, and highest arousal and number of errors occurred for anxiety-inducing musical content, followed by “happy” music, with “sad” music in between<span dir="RTL" lang="FA">.</span> This research has important implications for how music can affect behavioral safety systematically through physiological arousal, and how inter-individual differences are important to this process. Their work directly applies to developing intelligent in-car music recommendation systems, physiological detection technologies to prevent excessive arousal, and procedures for safe management of audio content, especially in challenging driving environments.This research investigates driving performance and physiological arousal as influenced by structural features of music using a within-subject design with 50 licensed participants. Fifteen music trials, created using fractional factorial design with factors including manipulated musical emotion, musical tempo, and volume, are conducted in the BeamNG-Drive simulator. The physiological measure SR was logged and scaled to create a common Arousal Index. The driving variables logged include driving errors, speeding pattern, lateral deviation from lanes, and reaction time<span dir="RTL" lang="FA">.</span> PCA on music-attitude data validated data adequacy, while k-means clustering extracted five distinct types of attitudes on music. Cluster analysis on features related to performance and physiology also extracted four stable patterns of responses from “low-error/low-arousal” to “high-error/high-arousal.” On within-subject statistical tests, musical condition had significant effects on both error rates and levels of arousal, and highest arousal and number of errors occurred for anxiety-inducing musical content, followed by “happy” music, with “sad” music in between<span dir="RTL" lang="FA">.</span> This research has important implications for how music can affect behavioral safety systematically through physiological arousal, and how inter-individual differences are important to this process. Their work directly applies to developing intelligent in-car music recommendation systems, physiological detection technologies to prevent excessive arousal, and procedures for safe management of audio content, especially in challenging driving environments.https://jte.sinaweb.net/article_236803_74d36491a242241fa16be110179563a8.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Traffic Safety Assessment in Combined Road and Land Use SectionsTraffic Safety Assessment in Combined Road and Land Use Sections5741575724540910.22119/jte.2026.562456.2754FAGhazalAhmadalinejadM.Sc., Department of Geotechnical and Transportation Engineering, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University (SBU), Tehran, IranMortezaAsadamrajiAssistant Professor, Department of Geotechnical and Transportation Engineering, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University (SBU), Tehran, Iran0000-0003-2799-4052Journal Article20251126Urban road design is in a reciprocal relationship with land use, the existence of each land use brings its own trip generation and attraction. Any change in land use characteristics can change the trip pattern. Optimal distribution of land uses can affect traffic safety of the land use area and the adjacent road. Multi-purpose land uses are considered important land uses due to their high trip attraction and generation. In this paper, the CODAS model has been used to evaluate the traffic safety effects of multi-purpose land uses. The CODAS model has the ability to combine and analyze different data. Also, the additional control performed in the CODAS method helps to increase the accuracy of modeling. From the combination of different land use groups and different functional categories, 20 cases of multi-purpose land uses in Tehran were selected and examined. After classifying the criteria related to land use and road, the Shannon entropy and best-worst combined methods have been used to weight the criteria. According to the results, the highest weight in the use section was assigned to "use infrastructure area" (0.324) and in the road section to "crash severity index" (0.226). Then, the road sections including the uses were ranked in terms of traffic safety with the combined CODAS model. This study is based on the combination of land uses. According to the results, multi-purpose land uses in the vicinity of the highway functional category with building group D have a higher safety risk, followed by group C in the vicinity of highway roads. The results of this study can be used in prioritizing and securing the area of multi-purpose uses, allocating safety budgets in the relevant area, and in implementing code 23 of the National Building Regulations in the field of traffic requirements for buildings.Urban road design is in a reciprocal relationship with land use, the existence of each land use brings its own trip generation and attraction. Any change in land use characteristics can change the trip pattern. Optimal distribution of land uses can affect traffic safety of the land use area and the adjacent road. Multi-purpose land uses are considered important land uses due to their high trip attraction and generation. In this paper, the CODAS model has been used to evaluate the traffic safety effects of multi-purpose land uses. The CODAS model has the ability to combine and analyze different data. Also, the additional control performed in the CODAS method helps to increase the accuracy of modeling. From the combination of different land use groups and different functional categories, 20 cases of multi-purpose land uses in Tehran were selected and examined. After classifying the criteria related to land use and road, the Shannon entropy and best-worst combined methods have been used to weight the criteria. According to the results, the highest weight in the use section was assigned to "use infrastructure area" (0.324) and in the road section to "crash severity index" (0.226). Then, the road sections including the uses were ranked in terms of traffic safety with the combined CODAS model. This study is based on the combination of land uses. According to the results, multi-purpose land uses in the vicinity of the highway functional category with building group D have a higher safety risk, followed by group C in the vicinity of highway roads. The results of this study can be used in prioritizing and securing the area of multi-purpose uses, allocating safety budgets in the relevant area, and in implementing code 23 of the National Building Regulations in the field of traffic requirements for buildings.https://jte.sinaweb.net/article_245409_ba2eafde95295b784d07b3a81080daf5.pdfParseh Designers Transportation Research InstituteQuarterly Journal of Transportation Engineering2008-659817420260622Causal Modeling of Factors Affecting the Performance of Autonomous Vehicles in Intelligent Transportation Systems (ITS)Causal Modeling of Factors Affecting the Performance of Autonomous Vehicles in Intelligent Transportation Systems (ITS)5759577124816710.22119/jte.2026.575462.2761FAElhamAsadamrajiAssistant Professor, Department of Information Technology Management, North Tehran Branch, Islamic Azad University, Tehran, Iran0000-0003-4453-9879Journal Article20260222Autonomous vehicles, as one of the key components of Intelligent Transportation Systems (ITS), play a significant role in enhancing the safety, efficiency, and sustainability of urban transportation networks. Despite considerable technological advancements in automation, sensor technologies, and communication systems, a comprehensive understanding of the causal relationships among the technical, human, infrastructural, and institutional factors influencing the performance of autonomous vehicles remains limited. The present study aims to develop and empirically test a comprehensive causal–statistical model within the ITS framework. The research data were collected from 320 experts and potential users through purposive and convenience sampling methods. Based on the 10-times rule in Partial Least Squares Structural Equation Modeling (PLS-SEM), the sample size was considered adequate for accurate parameter estimation. Data analysis was conducted using the PLS-SEM approach. The findings revealed that the level of autonomy, sensor reliability, and the quality of ITS communications have a positive and significant effect on transportation system efficiency through the mediating variables of trust, perceived safety, and technology acceptance. Furthermore, the results indicate that user trust and perceived safety play a decisive role in the acceptance and effective utilization of autonomous vehicles. Overall, the findings of this study provide a scientific and practical framework to support policymaking, infrastructure planning, and the successful development of intelligent transportation systems based on autonomous vehicles.Autonomous vehicles, as one of the key components of Intelligent Transportation Systems (ITS), play a significant role in enhancing the safety, efficiency, and sustainability of urban transportation networks. Despite considerable technological advancements in automation, sensor technologies, and communication systems, a comprehensive understanding of the causal relationships among the technical, human, infrastructural, and institutional factors influencing the performance of autonomous vehicles remains limited. The present study aims to develop and empirically test a comprehensive causal–statistical model within the ITS framework. The research data were collected from 320 experts and potential users through purposive and convenience sampling methods. Based on the 10-times rule in Partial Least Squares Structural Equation Modeling (PLS-SEM), the sample size was considered adequate for accurate parameter estimation. Data analysis was conducted using the PLS-SEM approach. The findings revealed that the level of autonomy, sensor reliability, and the quality of ITS communications have a positive and significant effect on transportation system efficiency through the mediating variables of trust, perceived safety, and technology acceptance. Furthermore, the results indicate that user trust and perceived safety play a decisive role in the acceptance and effective utilization of autonomous vehicles. Overall, the findings of this study provide a scientific and practical framework to support policymaking, infrastructure planning, and the successful development of intelligent transportation systems based on autonomous vehicles.https://jte.sinaweb.net/article_248167_aa1f76a9d323c6d1af8cdf2e53353c26.pdf