This paper proposes an iterative methodology to integrate large-scale behavioral activitybased models with mesoscopic traffic assignment models. The proposed approach fully decouples the two parts, allowing the ex-post integration of multiple models as long as certain assumptions are satisfied. A measure of error is defined to characterize a search space easily explorable within its boundaries. Within it, a joint distribution of the number of trips and travel times is identified as the equilibrium distribution, i.e., the distribution for which trip numbers and travel times are bound in the neighborhood of the equilibrium between supply and demand. The approach is tested on a medium-sized city of 400,000 inhabitants and the results suggest that the proposed iterative approach does perform well, reaching equilibrium between demand and supply in a limited number of iterations thanks to its perturbation techniques. Overall, 15 iterations are needed to reach values of the measure of error lower than 5%. The equilibrium identified this way is then validated against baseline distributions to demonstrate the goodness of the results.
{"title":"An Equilibrium-Seeking Search Algorithm for Integrating Large-Scale Activity-Based and Traffic Assignment Models","authors":"Serio Agriesti;Claudio Roncoli;Bat-Hen Nahmias-Biran","doi":"10.1109/OJITS.2025.3600918","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3600918","url":null,"abstract":"This paper proposes an iterative methodology to integrate large-scale behavioral activitybased models with mesoscopic traffic assignment models. The proposed approach fully decouples the two parts, allowing the ex-post integration of multiple models as long as certain assumptions are satisfied. A measure of error is defined to characterize a search space easily explorable within its boundaries. Within it, a joint distribution of the number of trips and travel times is identified as the equilibrium distribution, i.e., the distribution for which trip numbers and travel times are bound in the neighborhood of the equilibrium between supply and demand. The approach is tested on a medium-sized city of 400,000 inhabitants and the results suggest that the proposed iterative approach does perform well, reaching equilibrium between demand and supply in a limited number of iterations thanks to its perturbation techniques. Overall, 15 iterations are needed to reach values of the measure of error lower than 5%. The equilibrium identified this way is then validated against baseline distributions to demonstrate the goodness of the results.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1156-1170"},"PeriodicalIF":5.3,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11131172","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934536","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}
Pub Date : 2025-08-20DOI: 10.1109/OJITS.2025.3600966
Michael A. Tonkovich;Travis W. Moleski;Sam Fayez;Andrew Wallace;Preeti Choudhary;Jay P. Wilhelm
The development of autonomous driving technology has predominantly focused on urban and suburban areas. Deployment of automated driving systems in regions such as rural Appalachia present unique challenges such as narrow and winding roads and degradation of localization. Scenarios in rural Appalachia that required manual intervention by a driver during autonomous driving experiments were investigated across three unique routes. The research identified the technological and environmental limitations that contributed to these interventions and how they may differ from urban settings. The goal was to provide insights into the factors that hinder autonomous vehicle performance in rural areas and guide the development of more adaptable and robust systems capable of operating reliably in diverse environments, extending the benefits of autonomous driving to rural populations and ensuring equitable access to advancements in transportation. Driving experiments resulted in 1,884 total interventions and revealed trends in the reasons and locations for intervention across the three routes. In rural areas the leading causes of takeover were localization issues, accounting for 30.4% of total events, environmental traffic uncertainties, responsible for 20.3%, and object detection challenges, comprising 15.2%. Whereas urban settings saw roundabouts, environmental traffic uncertainties, and stoplight detection errors as the most common reasons with respective percentages of 19.5%, 17.7%, and 15.4%, revealing key differences between environments.
{"title":"Automated Driving System Challenges in Rural Appalachia","authors":"Michael A. Tonkovich;Travis W. Moleski;Sam Fayez;Andrew Wallace;Preeti Choudhary;Jay P. Wilhelm","doi":"10.1109/OJITS.2025.3600966","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3600966","url":null,"abstract":"The development of autonomous driving technology has predominantly focused on urban and suburban areas. Deployment of automated driving systems in regions such as rural Appalachia present unique challenges such as narrow and winding roads and degradation of localization. Scenarios in rural Appalachia that required manual intervention by a driver during autonomous driving experiments were investigated across three unique routes. The research identified the technological and environmental limitations that contributed to these interventions and how they may differ from urban settings. The goal was to provide insights into the factors that hinder autonomous vehicle performance in rural areas and guide the development of more adaptable and robust systems capable of operating reliably in diverse environments, extending the benefits of autonomous driving to rural populations and ensuring equitable access to advancements in transportation. Driving experiments resulted in 1,884 total interventions and revealed trends in the reasons and locations for intervention across the three routes. In rural areas the leading causes of takeover were localization issues, accounting for 30.4% of total events, environmental traffic uncertainties, responsible for 20.3%, and object detection challenges, comprising 15.2%. Whereas urban settings saw roundabouts, environmental traffic uncertainties, and stoplight detection errors as the most common reasons with respective percentages of 19.5%, 17.7%, and 15.4%, revealing key differences between environments.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1121-1132"},"PeriodicalIF":5.3,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11131208","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934412","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}
Pub Date : 2025-08-20DOI: 10.1109/OJITS.2025.3600667
Xue Xing;Fahui Luo;Bin Wang;Yufei Huang;Lei Tang
The study introduces a new method to enhance vehicle type recognition rates in Internet of Vehicles environment. The approach integrates a vehicle target detection model that utilizes bidirectional feature fusion of a hybrid attention mechanism and an enhanced CenterNet encoding technique, with ResNet18 as the base network. By decoupling detection and classification processes, the model focuses on vehicle characteristics and unique model differences, boosting accuracy. Additionally, Scale feature information is incorporated to improve CenterNet vehicle target detection by learning width, height, and shape details. To address low detection rates of specific vehicle models like buses and vans, a bidirectional feature fusion mechanism is employed, combining a hybrid attention mechanism (CBAMBiFPN) to enhance feature utilization and detection accuracy. Experimental results on UA-DETRAC and BDD datasets demonstrated an average accuracy increase, validating the model’s effectiveness. Compared to the original model, the new model showed improvements in mean average precision, F1-Score, and detection speed. Specifically, the UA-DETRAC data set saw a 1.6 percentage point increase in mean average precision and a 1.8 percentage point increase in F1-Score, with a detection speed of 68 frames/s. On the BDD100K data set, the model improved mean average precision by 1.1 percentage points. The study showcases enhanced accuracy without compromising real-time performance.
{"title":"Vehicle Target Detection Model Based on CBAM-BiFPN and Improved CenterNet Coding","authors":"Xue Xing;Fahui Luo;Bin Wang;Yufei Huang;Lei Tang","doi":"10.1109/OJITS.2025.3600667","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3600667","url":null,"abstract":"The study introduces a new method to enhance vehicle type recognition rates in Internet of Vehicles environment. The approach integrates a vehicle target detection model that utilizes bidirectional feature fusion of a hybrid attention mechanism and an enhanced CenterNet encoding technique, with ResNet18 as the base network. By decoupling detection and classification processes, the model focuses on vehicle characteristics and unique model differences, boosting accuracy. Additionally, Scale feature information is incorporated to improve CenterNet vehicle target detection by learning width, height, and shape details. To address low detection rates of specific vehicle models like buses and vans, a bidirectional feature fusion mechanism is employed, combining a hybrid attention mechanism (CBAMBiFPN) to enhance feature utilization and detection accuracy. Experimental results on UA-DETRAC and BDD datasets demonstrated an average accuracy increase, validating the model’s effectiveness. Compared to the original model, the new model showed improvements in mean average precision, F1-Score, and detection speed. Specifically, the UA-DETRAC data set saw a 1.6 percentage point increase in mean average precision and a 1.8 percentage point increase in F1-Score, with a detection speed of 68 frames/s. On the BDD100K data set, the model improved mean average precision by 1.1 percentage points. The study showcases enhanced accuracy without compromising real-time performance.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1143-1155"},"PeriodicalIF":5.3,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11131167","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144990064","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}
Pub Date : 2025-08-19DOI: 10.1109/OJITS.2025.3600482
Zhijun Fu;Beibei Chai;Dengfeng Zhao;Bao Ma;Subhash Rakheja;Jia Hu
This study addresses the limitations of existing collaborative control systems for mixed traffic environments where connected and automated vehicles (CAVs) coexist with human-driven vehicles (HDVs), which overemphasize functional safety and energy efficiency loops while neglecting comfort. We propose a hierarchical model predictive control (HMPC) framework incorporating occupants’ motion sickness. The upper layer generates globally optimal speed sequences through dynamic prediction of signal phases, while the lower layer adopts a variable-weight MPC optimization method with a composite cost function integrating travel time, delay, and motion sickness indicators. To address varying CAV penetration rates in mixed traffic, heterogeneous vehicle dynamics models are developed, where CAVs and HDVs employ Cooperative Adaptive Cruise Control (CACC) and Intelligent Driver Model (IDM), respectively. The simulation evaluation results demonstrates that the proposed method achieves significant performance enhancements across diverse CAV penetration rates and traffic saturation scenarios: traffic efficiency is improved by 6.30% and 13.94%, while motion comfort is improved by 51.91% and 25.07%. Field evaluation at the Dongfeng-Huayuan Road intersection in Zhengzhou further confirms these findings, showing 28.97% and 37.87% reductions in travel time and delay, together with 57.81% and 18.18% declines in MSDV and RMS-Jerk, thereby confirming the control strategy’s robustness in real-world perturbed environments.
{"title":"Motion Sickness-Oriented Cooperative Control in Mixed Traffic: A Hierarchical MPC Framework With Multi-Objective Optimization","authors":"Zhijun Fu;Beibei Chai;Dengfeng Zhao;Bao Ma;Subhash Rakheja;Jia Hu","doi":"10.1109/OJITS.2025.3600482","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3600482","url":null,"abstract":"This study addresses the limitations of existing collaborative control systems for mixed traffic environments where connected and automated vehicles (CAVs) coexist with human-driven vehicles (HDVs), which overemphasize functional safety and energy efficiency loops while neglecting comfort. We propose a hierarchical model predictive control (HMPC) framework incorporating occupants’ motion sickness. The upper layer generates globally optimal speed sequences through dynamic prediction of signal phases, while the lower layer adopts a variable-weight MPC optimization method with a composite cost function integrating travel time, delay, and motion sickness indicators. To address varying CAV penetration rates in mixed traffic, heterogeneous vehicle dynamics models are developed, where CAVs and HDVs employ Cooperative Adaptive Cruise Control (CACC) and Intelligent Driver Model (IDM), respectively. The simulation evaluation results demonstrates that the proposed method achieves significant performance enhancements across diverse CAV penetration rates and traffic saturation scenarios: traffic efficiency is improved by 6.30% and 13.94%, while motion comfort is improved by 51.91% and 25.07%. Field evaluation at the Dongfeng-Huayuan Road intersection in Zhengzhou further confirms these findings, showing 28.97% and 37.87% reductions in travel time and delay, together with 57.81% and 18.18% declines in MSDV and RMS-Jerk, thereby confirming the control strategy’s robustness in real-world perturbed environments.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1133-1142"},"PeriodicalIF":5.3,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11130523","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144934501","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}
The adoption of Electric Vehicles (EVs) is a transformative step toward reducing CO2 emissions and achieving global sustainability targets such as the UN Sustainable Development Goals and IMT-2030. However, large-scale EV integration poses significant challenges across multiple interdependent domains: the Power Grid (PG), Transportation Systems (TS), and Information and Communication Technology (ICT). Most existing research approaches these sectors independently, lacking a holistic view of their interconnected constraints and resource dependencies. This review presents a comprehensive and critical analysis of cross-sector resource management for EV integration, emphasizing the interrelations among PG, TS, and ICT. We identify key resource requirements, examine cross-domain challenges, and evaluate the effectiveness of current solutions and Key-Enabling Technologies (KETs) in addressing them. Through this analysis, we highlight critical research gaps and advocate for a unified and collaborative crosssector approach to resource management. By offering this cross-sector resource management perspective, the review contributes new insights to guide future research and policy development in support of sustainable and scalable EV ecosystem integration aligned with 6G and IMT-2030 visions.
{"title":"A Review on the Cross-Sector Resource Management Framework for Electric Vehicles Integration: Challenges, Solutions, Key-Enabling Technologies, and Future Directions","authors":"Narges Gholipoor;Mehdi Rasti;Fahimeh Aghaei;Farid Hamzeh Aghdam;Abdelhak Kharbouch;Valiollah Talaeizadeh;Jamshid Aghaei;Hesham A. Rakha","doi":"10.1109/OJITS.2025.3593437","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3593437","url":null,"abstract":"The adoption of Electric Vehicles (EVs) is a transformative step toward reducing CO2 emissions and achieving global sustainability targets such as the UN Sustainable Development Goals and IMT-2030. However, large-scale EV integration poses significant challenges across multiple interdependent domains: the Power Grid (PG), Transportation Systems (TS), and Information and Communication Technology (ICT). Most existing research approaches these sectors independently, lacking a holistic view of their interconnected constraints and resource dependencies. This review presents a comprehensive and critical analysis of cross-sector resource management for EV integration, emphasizing the interrelations among PG, TS, and ICT. We identify key resource requirements, examine cross-domain challenges, and evaluate the effectiveness of current solutions and Key-Enabling Technologies (KETs) in addressing them. Through this analysis, we highlight critical research gaps and advocate for a unified and collaborative crosssector approach to resource management. By offering this cross-sector resource management perspective, the review contributes new insights to guide future research and policy development in support of sustainable and scalable EV ecosystem integration aligned with 6G and IMT-2030 visions.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1084-1120"},"PeriodicalIF":5.3,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11099546","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904751","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}
Pub Date : 2025-07-25DOI: 10.1109/OJITS.2025.3592628
Tsung-Han Tsai;Wei-Chung Wan
Monocular depth estimation is an important topic in computer vision. Recently the CNNs (Convolutional Neural Networks) based model shows a reasonable result from an end-to-end encoder-decoder architecture. For the encoder part, most of the research is based on a robust feature extractor to get good features. With a strong encoder, it was found that even simple up-sampling processes can achieve good accuracy. However, the decoder part is more critical in a high-quality depth estimation task. Even now, there is no intuitive way to calibrate the feature map for the upsampling process. In this paper, we present a novel monocular depth estimation design. We propose an innovative CNN-based network module that considers the whole up-sampling process globally. This design is based on the concept of SE-Net, and properly recalibrated the feature maps with a global perspective attention mechanism. We further combine it with Non-local network attention mechanisms to design the Non-Local Decoder-Squeeze-and-Excitation (NL-DSE) module for the whole up-sampling process. Furthermore, we also propose an output limiting range method called Adaptive Depth List (ADL) to enhance the precision of the near-distance estimation. Combining these proposed techniques, our results are evaluated on the NYU Depth V2 dataset and outperform the state-of-the-art CNN-based approaches in accuracy.
{"title":"Monocular Depth Estimation by Non-Local Decoder-Squeeze-and-Excitation Network With Adaptive Depth List","authors":"Tsung-Han Tsai;Wei-Chung Wan","doi":"10.1109/OJITS.2025.3592628","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3592628","url":null,"abstract":"Monocular depth estimation is an important topic in computer vision. Recently the CNNs (Convolutional Neural Networks) based model shows a reasonable result from an end-to-end encoder-decoder architecture. For the encoder part, most of the research is based on a robust feature extractor to get good features. With a strong encoder, it was found that even simple up-sampling processes can achieve good accuracy. However, the decoder part is more critical in a high-quality depth estimation task. Even now, there is no intuitive way to calibrate the feature map for the upsampling process. In this paper, we present a novel monocular depth estimation design. We propose an innovative CNN-based network module that considers the whole up-sampling process globally. This design is based on the concept of SE-Net, and properly recalibrated the feature maps with a global perspective attention mechanism. We further combine it with Non-local network attention mechanisms to design the Non-Local Decoder-Squeeze-and-Excitation (NL-DSE) module for the whole up-sampling process. Furthermore, we also propose an output limiting range method called Adaptive Depth List (ADL) to enhance the precision of the near-distance estimation. Combining these proposed techniques, our results are evaluated on the NYU Depth V2 dataset and outperform the state-of-the-art CNN-based approaches in accuracy.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1069-1083"},"PeriodicalIF":5.3,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11096588","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144887810","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}
This paper proposes a simple yet unexplored measurement and federated learning system architecture for connected vehicles. The novelty of the introduced system is to combine the real-time data-sharing of crowdsensing with federated learning of global traffic models, providing up-to-date information for decision-making, facilitating faster learning, improving communicational channel usage, and possibly enhancing data privacy. This multi-level cooperative federated learning system generally supports operational, tactical, and strategic planning; therefore, we demonstrate its merits with a case study of parking monitoring in a simulated town as well as average speed prediction in a simulated part of Hannover, Germany. However, real-time data-sharing is essential for decision-making; it might also contain privacy-sensitive information regarding the trajectory of the vehicles. To mitigate the risk of privacy leakage, we experimented with different data selection methods for data exchange, introducing an optimization method inspired by Zeuthen’s negotiation strategy. We also checked the privacy impact of real-time data-sharing on federated learning. Our results indicate only negligible differences in privacy leakage between the proposed data selection methods. On the other hand, real-time data-sharing improves the reaction time of the federated learning system. The Zeuthen-inspired optimization method can efficiently supply valuable information for the communication partners. Moreover, it can enhance privacy protection in federated learning in some cases.
{"title":"Introducing Intelligent Data Sharing to Vehicular Cooperative Federated Learning","authors":"Levente Alekszejenkó;Péter Antal;Tadeusz Dobrowiecki","doi":"10.1109/OJITS.2025.3589612","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3589612","url":null,"abstract":"This paper proposes a simple yet unexplored measurement and federated learning system architecture for connected vehicles. The novelty of the introduced system is to combine the real-time data-sharing of crowdsensing with federated learning of global traffic models, providing up-to-date information for decision-making, facilitating faster learning, improving communicational channel usage, and possibly enhancing data privacy. This multi-level cooperative federated learning system generally supports operational, tactical, and strategic planning; therefore, we demonstrate its merits with a case study of parking monitoring in a simulated town as well as average speed prediction in a simulated part of Hannover, Germany. However, real-time data-sharing is essential for decision-making; it might also contain privacy-sensitive information regarding the trajectory of the vehicles. To mitigate the risk of privacy leakage, we experimented with different data selection methods for data exchange, introducing an optimization method inspired by Zeuthen’s negotiation strategy. We also checked the privacy impact of real-time data-sharing on federated learning. Our results indicate only negligible differences in privacy leakage between the proposed data selection methods. On the other hand, real-time data-sharing improves the reaction time of the federated learning system. The Zeuthen-inspired optimization method can efficiently supply valuable information for the communication partners. Moreover, it can enhance privacy protection in federated learning in some cases.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1009-1026"},"PeriodicalIF":5.3,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11082276","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144758451","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}
Pub Date : 2025-07-14DOI: 10.1109/OJITS.2025.3589208
Harshit Maheshwari;Li Yang;Richard W. Pazzi
Urban traffic simulation is useful in many ways to understand, manage, and predict the growing complexities of traffic dynamics within a city. Traditional simulation models often struggle to capture the intricacies of urban traffic patterns, leading to unrealistic simulations, which negatively affect traffic management and urban planning. In recent years, Machine Learning solutions have emerged to enhance various aspects of urban traffic simulation, which is possible by utilizing vast amounts of data and extracting valuable insights. This survey systematically reviews the state-of-the-art Machine Learning techniques applied to urban traffic simulation. By focusing on the practical application of Machine Learning techniques in various studies, we aim to analyze the current research direction, highlight the effectiveness of existing approaches, identify their limitations, and propose potential strategies to improve the performance and applicability of these techniques in real-world scenarios. Another key contribution of this survey is a proof-of-concept case study, which utilizes a basic Reinforcement Learning algorithm to control traffic lights across multiple intersections. The results from this case study demonstrate a significant improvement in vehicle wait time compared to the static baseline method. The code developed for this case study is publicly available, providing a valuable resource for researchers interested in replicating this work or building upon it. This survey aims to bridge the gap between simulation and reality by providing a comprehensive foundational understanding of the subject, critically evaluating the existing limitations in current methodologies, and suggesting future directions to improve performance, adaptability, and usability.
{"title":"Machine Learning Advancements in Urban Traffic Simulation: A Comprehensive Survey","authors":"Harshit Maheshwari;Li Yang;Richard W. Pazzi","doi":"10.1109/OJITS.2025.3589208","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3589208","url":null,"abstract":"Urban traffic simulation is useful in many ways to understand, manage, and predict the growing complexities of traffic dynamics within a city. Traditional simulation models often struggle to capture the intricacies of urban traffic patterns, leading to unrealistic simulations, which negatively affect traffic management and urban planning. In recent years, Machine Learning solutions have emerged to enhance various aspects of urban traffic simulation, which is possible by utilizing vast amounts of data and extracting valuable insights. This survey systematically reviews the state-of-the-art Machine Learning techniques applied to urban traffic simulation. By focusing on the practical application of Machine Learning techniques in various studies, we aim to analyze the current research direction, highlight the effectiveness of existing approaches, identify their limitations, and propose potential strategies to improve the performance and applicability of these techniques in real-world scenarios. Another key contribution of this survey is a proof-of-concept case study, which utilizes a basic Reinforcement Learning algorithm to control traffic lights across multiple intersections. The results from this case study demonstrate a significant improvement in vehicle wait time compared to the static baseline method. The code developed for this case study is publicly available, providing a valuable resource for researchers interested in replicating this work or building upon it. This survey aims to bridge the gap between simulation and reality by providing a comprehensive foundational understanding of the subject, critically evaluating the existing limitations in current methodologies, and suggesting future directions to improve performance, adaptability, and usability.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"1027-1052"},"PeriodicalIF":5.3,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11079941","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144758429","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}
Efficient parking management is crucial in crowded Asian cities to optimize limited road space and parking facilities. The increasing vehicle ownership rate in Taiwan has intensified the demand for street parking, leading to excessive driving in search of available spots and contributing up to 30% of traffic congestion. This paper proposes a low-cost, infrastructure-free outdoor roadside parking management system based on high-definition (HD) map updating. The system fuses data from a solid-state LiDAR (SSL) system, a monocular camera, an inertial navigation system, a GPS, and HD maps followed by deep-learning-based efficient region extraction. The goal was to achieve high accuracy with minimal computational resources and infrastructure costs. The proposed system’s performance for dynamic HD map object updating was evaluated through parking management tests. The system’s costs were low due to the selection of SSL and monocular cameras. Traditional and novel extrinsic calibration methods were compared in various experiments, and a hardware architecture for precise sensor time synchronization was designed. Software algorithms for accurate image–point-cloud projection were developed to update HD map parking layers. By using normal distribution transform matching of the SSL and HD point cloud maps, navigation performance was achieved to 0.4-meter accuracy. When applied to license plate localization in two experimental scenarios, the mean performance error was approximately 0.48 and 0.62 m.
{"title":"AI-Driven Mapping System for Smart Parking Management Applications Using an INS-GNSS-Solid-State LiDAR-Monocular Camera Fusion Engine Empowered by HD Maps","authors":"Kai-Wei Chiang;Syun Tsai;Jou-An Chen;Surachet Srinara;Meng-Lun Tsai;Chih-Yun Hsieh;Jyun-Yang Hung;Chalermchon Satirapod;Naser El-Sheimy","doi":"10.1109/OJITS.2025.3587274","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3587274","url":null,"abstract":"Efficient parking management is crucial in crowded Asian cities to optimize limited road space and parking facilities. The increasing vehicle ownership rate in Taiwan has intensified the demand for street parking, leading to excessive driving in search of available spots and contributing up to 30% of traffic congestion. This paper proposes a low-cost, infrastructure-free outdoor roadside parking management system based on high-definition (HD) map updating. The system fuses data from a solid-state LiDAR (SSL) system, a monocular camera, an inertial navigation system, a GPS, and HD maps followed by deep-learning-based efficient region extraction. The goal was to achieve high accuracy with minimal computational resources and infrastructure costs. The proposed system’s performance for dynamic HD map object updating was evaluated through parking management tests. The system’s costs were low due to the selection of SSL and monocular cameras. Traditional and novel extrinsic calibration methods were compared in various experiments, and a hardware architecture for precise sensor time synchronization was designed. Software algorithms for accurate image–point-cloud projection were developed to update HD map parking layers. By using normal distribution transform matching of the SSL and HD point cloud maps, navigation performance was achieved to 0.4-meter accuracy. When applied to license plate localization in two experimental scenarios, the mean performance error was approximately 0.48 and 0.62 m.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"995-1008"},"PeriodicalIF":5.3,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11075854","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144725186","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}
Pub Date : 2025-07-04DOI: 10.1109/OJITS.2025.3581617
Guanyu Lin;Sean Qian;Zulqarnain H. Khattak
The increasing prevalence of connected and autonomous vehicles (CAVs) in smart cities requires robust cyberattack and anomaly detection systems to ensure safety and resilience. Cyberattacks on leader and follower in cooperative driving can result in differing impacts, however, their impacts on security and resilience of cooperative driving are largely unknown. Traditional anomaly detection methods, which aggregate data centrally, compromise driver privacy and are insufficient to address real-world challenges due to limitations of being compromised by adversarial attacks. To overcome these limitations, we propose Explainable Fine-Grained Cyberattacks and Anomaly Detection with Federated Agents for connected autonomous vehicles (xFedCAV). Our framework leverages federated learning to enhance privacy and security, using Shapley Additive exPlanations (SHAP) for interpretable detection. Unlike existing methods, xFedCAV focuses on fine-grained detection by simulating cyberattacks on individual vehicles rather than the entire fleet, allowing for more precise identification and response. Experimental results, conducted on a real-world CAV dataset, demonstrate that xFedCAV not only explains the relationship between vehicle characteristics and detection outputs, but also effectively detects cyberattacks in a decentralized manner. This research offers knowledge about the cybersecurity impacts of the leader and follower within cooperative driving and provides a significant advancement in federated learning applications for CAVs, contributing to the development of safer and more resilient smart city applications for transportation systems.
{"title":"xFedCAV: Cyberattacks on Leader and Followers in Automated Vehicles With Cooperative Platoons Using Federated Agents","authors":"Guanyu Lin;Sean Qian;Zulqarnain H. Khattak","doi":"10.1109/OJITS.2025.3581617","DOIUrl":"https://doi.org/10.1109/OJITS.2025.3581617","url":null,"abstract":"The increasing prevalence of connected and autonomous vehicles (CAVs) in smart cities requires robust cyberattack and anomaly detection systems to ensure safety and resilience. Cyberattacks on leader and follower in cooperative driving can result in differing impacts, however, their impacts on security and resilience of cooperative driving are largely unknown. Traditional anomaly detection methods, which aggregate data centrally, compromise driver privacy and are insufficient to address real-world challenges due to limitations of being compromised by adversarial attacks. To overcome these limitations, we propose Explainable Fine-Grained Cyberattacks and Anomaly Detection with Federated Agents for connected autonomous vehicles (xFedCAV). Our framework leverages federated learning to enhance privacy and security, using Shapley Additive exPlanations (SHAP) for interpretable detection. Unlike existing methods, xFedCAV focuses on fine-grained detection by simulating cyberattacks on individual vehicles rather than the entire fleet, allowing for more precise identification and response. Experimental results, conducted on a real-world CAV dataset, demonstrate that xFedCAV not only explains the relationship between vehicle characteristics and detection outputs, but also effectively detects cyberattacks in a decentralized manner. This research offers knowledge about the cybersecurity impacts of the leader and follower within cooperative driving and provides a significant advancement in federated learning applications for CAVs, contributing to the development of safer and more resilient smart city applications for transportation systems.","PeriodicalId":100631,"journal":{"name":"IEEE Open Journal of Intelligent Transportation Systems","volume":"6 ","pages":"898-914"},"PeriodicalIF":4.6,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11071968","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144646590","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：