Recent advances in autonomous system simulation platforms have significantly enhanced the safe and scalable testing of driving policies. Although existing simulators have greatly accelerated development by providing controlled testing environments, they face limitations in addressing the evolving needs of future transportation research, particularly in enabling effective human-artificial intelligence (human-AI) collaboration and modeling socially aware driving agents. This study introduces Sky-Drive, a novel distributed multiagent simulation platform that addresses these limitations through four key innovations: (1) a distributed architecture for synchronized simulation across multiple terminals; (2) a multimodal human-in-the-loop framework that integrates diverse sensors to collect rich behavioral data; (3) a human-AI collaboration mechanism that supports continuous and adaptive knowledge exchange; and (4) a digital twin framework for constructing high-fidelity virtual replicas of real-world transportation environments. Sky-Drive supports diverse applications, such as autonomous vehicle-human road user interaction modeling, human-in-the-loop training, socially aware reinforcement learning, personalized driving development, and customized scenario generation. Future extensions will incorporate foundation models for context-aware decision support and hardware-in-the-loop testing for real-world validation. By bridging scenario generation, data collection, algorithm training, and hardware integration, Sky-Drive has the potential to become a foundational platform for the next generation of human-centered and socially aware autonomous transportation system research.
{"title":"Sky-Drive: a Distributed Multiagent Simulation Platform for Human-AI Collaborative and Socially Aware Future Transportation","authors":"Zilin Huang;Zihao Sheng;Zhengyang Wan;Yansong Qu;Yuhao Luo;Boyue Wang;Pei Li;Yen-Jung Chen;Jiancong Chen;Keke Long;Jiayi Meng;Yue Leng;Sikai Chen","doi":"10.26599/JICV.2026.9210070","DOIUrl":"https://doi.org/10.26599/JICV.2026.9210070","url":null,"abstract":"Recent advances in autonomous system simulation platforms have significantly enhanced the safe and scalable testing of driving policies. Although existing simulators have greatly accelerated development by providing controlled testing environments, they face limitations in addressing the evolving needs of future transportation research, particularly in enabling effective human-artificial intelligence (human-AI) collaboration and modeling socially aware driving agents. This study introduces Sky-Drive, a novel distributed multiagent simulation platform that addresses these limitations through four key innovations: (1) a distributed architecture for synchronized simulation across multiple terminals; (2) a multimodal human-in-the-loop framework that integrates diverse sensors to collect rich behavioral data; (3) a human-AI collaboration mechanism that supports continuous and adaptive knowledge exchange; and (4) a digital twin framework for constructing high-fidelity virtual replicas of real-world transportation environments. Sky-Drive supports diverse applications, such as autonomous vehicle-human road user interaction modeling, human-in-the-loop training, socially aware reinforcement learning, personalized driving development, and customized scenario generation. Future extensions will incorporate foundation models for context-aware decision support and hardware-in-the-loop testing for real-world validation. By bridging scenario generation, data collection, algorithm training, and hardware integration, Sky-Drive has the potential to become a foundational platform for the next generation of human-centered and socially aware autonomous transportation system research.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210070-1-9210070-16"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319351","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852551","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 : 2025-12-30DOI: 10.26599/JICV.2026.9210068
Tianpei Tang;Xiaofan Xue;Shengnan Zhao;Bang Luo;Hua Wang
The successful deployment of autonomous vehicles (AVs) relies heavily on their ability to interact safely and effectively with other road users. External human-machine interfaces (eHMIs) have emerged as critical components in facilitating these interactions. Rigorous evaluation and testing of eHMIs are essential for realizing their intended safety and communication benefits. This study provides a comprehensive review of current eHMI research, focusing on their impact on road users' behavior and perceptions, as well as the methods used for evaluation. Key behavioral factors-such as eHMI modality, information type, location, vehicle kinematics, traffic environment, and user characteristics-are systematically reviewed and summarized. The influence of eHMIs on user perceptions is also explored through indicators such as perceived safety, comprehensibility, trust, cognitive load, and user experience. Despite notable advancements, several critical research gaps remain underexplored. Most studies focus on one-to-one interactions, neglecting the complexities of mixed-traffic environments involving AVs, conventional human-driven vehicles, pedestrians, and cyclists. Current evaluation methods largely rely on virtual reality and Wizard-of-Oz experiments, which may fail to fully capture real-world dynamics. Additionally, subjective questionnaires, which are often used in these studies, do not guarantee high reproducibility of findings. Moreover, insufficient attention has been given to the synchronization of eHMI signals with vehicle kinematics. Furthermore, the absence of standardized evaluation frameworks limits cross-study comparability and the development of universally applicable eHMI solutions. To address these challenges, future research should prioritize the integration of naturalistic traffic scenarios, the adoption of objective and reproducible evaluation methods, the exploration of multimodal eHMI designs, and the development of standardized assessment protocols. These efforts are crucial for improving AV communication with diverse road users and ensuring safety in increasingly complex traffic ecosystems.
{"title":"Road Users' Behavior and Perceptions of Autonomous Vehicles with External Human-Machine Interfaces: a Review of Developments in 2017–2024","authors":"Tianpei Tang;Xiaofan Xue;Shengnan Zhao;Bang Luo;Hua Wang","doi":"10.26599/JICV.2026.9210068","DOIUrl":"https://doi.org/10.26599/JICV.2026.9210068","url":null,"abstract":"The successful deployment of autonomous vehicles (AVs) relies heavily on their ability to interact safely and effectively with other road users. External human-machine interfaces (eHMIs) have emerged as critical components in facilitating these interactions. Rigorous evaluation and testing of eHMIs are essential for realizing their intended safety and communication benefits. This study provides a comprehensive review of current eHMI research, focusing on their impact on road users' behavior and perceptions, as well as the methods used for evaluation. Key behavioral factors-such as eHMI modality, information type, location, vehicle kinematics, traffic environment, and user characteristics-are systematically reviewed and summarized. The influence of eHMIs on user perceptions is also explored through indicators such as perceived safety, comprehensibility, trust, cognitive load, and user experience. Despite notable advancements, several critical research gaps remain underexplored. Most studies focus on one-to-one interactions, neglecting the complexities of mixed-traffic environments involving AVs, conventional human-driven vehicles, pedestrians, and cyclists. Current evaluation methods largely rely on virtual reality and Wizard-of-Oz experiments, which may fail to fully capture real-world dynamics. Additionally, subjective questionnaires, which are often used in these studies, do not guarantee high reproducibility of findings. Moreover, insufficient attention has been given to the synchronization of eHMI signals with vehicle kinematics. Furthermore, the absence of standardized evaluation frameworks limits cross-study comparability and the development of universally applicable eHMI solutions. To address these challenges, future research should prioritize the integration of naturalistic traffic scenarios, the adoption of objective and reproducible evaluation methods, the exploration of multimodal eHMI designs, and the development of standardized assessment protocols. These efforts are crucial for improving AV communication with diverse road users and ensuring safety in increasingly complex traffic ecosystems.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210068-1-9210068-18"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319349","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852559","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}
Autonomous driving technology is becoming increasingly popular, transforming transportation systems worldwide. However, its perception modules are highly vulnerable to adversarial attacks, which exploit weaknesses in deep neural networks, leading to potential safety risks and compromised decision-making in autonomous systems. In this study, we propose AdvGLOW, a novel adversarial attack model tailored for covert attacks on autonomous driving perception modules in traffic scenarios. Leveraging an information exchange network within a flow-based model, AdvGLOW introduces reversible data transformations to achieve high attack success with minimal perturbation visibility. By optimizing a combined global-local loss, our model preserves structural details while embedding adversarial features, resulting in robust yet visually imperceptible adversarial samples. We conduct extensive experiments on traffic-related datasets, demonstrating that the generated adversarial samples are challenging for both humans and algorithms to detect. Additionally, this method exhibits strong attack robustness and transferability.
{"title":"AdvGLOW: Covert Adversarial Attacks Against Autonomous Driving Perception","authors":"Xuesong Bai;Peng Dong;Jinlei Wang;Yuanhao Huang;Haiyang Yu;Yilong Ren","doi":"10.26599/JICV.2025.9210067","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210067","url":null,"abstract":"Autonomous driving technology is becoming increasingly popular, transforming transportation systems worldwide. However, its perception modules are highly vulnerable to adversarial attacks, which exploit weaknesses in deep neural networks, leading to potential safety risks and compromised decision-making in autonomous systems. In this study, we propose AdvGLOW, a novel adversarial attack model tailored for covert attacks on autonomous driving perception modules in traffic scenarios. Leveraging an information exchange network within a flow-based model, AdvGLOW introduces reversible data transformations to achieve high attack success with minimal perturbation visibility. By optimizing a combined global-local loss, our model preserves structural details while embedding adversarial features, resulting in robust yet visually imperceptible adversarial samples. We conduct extensive experiments on traffic-related datasets, demonstrating that the generated adversarial samples are challenging for both humans and algorithms to detect. Additionally, this method exhibits strong attack robustness and transferability.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210067-1-9210067-13"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319353","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852469","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 : 2025-12-30DOI: 10.26599/JICV.2026.9210072
Wenrui Jin;Xiaoxiao Lv;Xiangping Qiu;Fan Mo;Min Fang;Jiaxue Li
The coexistence of manual and autonomous driving leads to a nonstandard final parking state (FPS) of adjacent vehicles. In these instances, maintaining a standard FPS for the target vehicle will engender a predicament of challenging door access. To increase the comfort of passenger boarding and alighting (CPBA), this study proposes a bidirectional search hybrid A* (BHA*) method for parking path planning. First, a characterization variable for the CPBA is constructed on the basis of the allowable and required opening angles of vehicle doors under the constraint of adjacent vehicles. An optimization model for FPS is subsequently established with the comprehensive objective of the CPBA for both the target and adjacent vehicles, along with a safe distance. The genetic algorithm is then utilized to obtain the optimal FPS. Furthermore, the Voronoi potential and a bidirectional search strategy are employed to improve the hybrid A* algorithm, aiming to achieve the optimal FPS and enhance the efficiency of parking path planning. Finally, simulation experiments are conducted on the parameters of real vehicles and parking scenarios to verify the effectiveness and adaptability of the proposed method. A comparison with the hybrid A* algorithm further confirms the superiority of the proposed method in terms of search efficiency.
{"title":"An Efficient Bidirectional Search Hybrid A* Method for Parking Path Planning in Adjacent Vehicle Deviation Scenarios to Enhance Passenger Comfort","authors":"Wenrui Jin;Xiaoxiao Lv;Xiangping Qiu;Fan Mo;Min Fang;Jiaxue Li","doi":"10.26599/JICV.2026.9210072","DOIUrl":"https://doi.org/10.26599/JICV.2026.9210072","url":null,"abstract":"The coexistence of manual and autonomous driving leads to a nonstandard final parking state (FPS) of adjacent vehicles. In these instances, maintaining a standard FPS for the target vehicle will engender a predicament of challenging door access. To increase the comfort of passenger boarding and alighting (CPBA), this study proposes a bidirectional search hybrid A* (BHA*) method for parking path planning. First, a characterization variable for the CPBA is constructed on the basis of the allowable and required opening angles of vehicle doors under the constraint of adjacent vehicles. An optimization model for FPS is subsequently established with the comprehensive objective of the CPBA for both the target and adjacent vehicles, along with a safe distance. The genetic algorithm is then utilized to obtain the optimal FPS. Furthermore, the Voronoi potential and a bidirectional search strategy are employed to improve the hybrid A* algorithm, aiming to achieve the optimal FPS and enhance the efficiency of parking path planning. Finally, simulation experiments are conducted on the parameters of real vehicles and parking scenarios to verify the effectiveness and adaptability of the proposed method. A comparison with the hybrid A* algorithm further confirms the superiority of the proposed method in terms of search efficiency.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210072-1-9210072-16"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319352","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852526","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 : 2025-12-30DOI: 10.26599/JICV.2026.9210069
Yu Yan;Chengxi Zhang;Jin Wu;Chen Sun;Qingwen Meng;Chaoyi Chen;Jianqiang Wang;Guangwei Wang
Autonomous driving on mountainous roads is challenging since numerous curves and varying road slopes make vehicle control difficult. To address this problem, we propose triple adaptive delayed deep deterministic (TAD3) policy gradient, a novel end-to-end adaptive deep reinforcement learning (DRL) system based on the twin delayed deep deterministic policy gradient (TD3) algorithm. This approach combines recurrent neural networks (RNNs) with the critic network to leverage the vehicle's historical state information and incorporates a triple-critics structure to adapt to diverse mountainous road conditions with various curvatures and gradients. Compared with previous methods, the proposed TAD3 approach achieves better vehicle performance on mountain roads and faster training speeds without extensive knowledge of vehicle dynamics. The experimental results using the TORCS simulator demonstrate that the proposed TAD3 achieves a 43%-64% smaller distance error and a 19%-74% smaller yaw angle error than five state-of-the-art baselines do in the lane-keeping task while simultaneously achieving lower laptimes in the time-minimum task and demonstrating superior generalization ability on three mountainous tracks with different designs in terms of curvature and gradient.
{"title":"Autonomous Driving on Mountain Roads via an Adaptive Deep Reinforcement Learning Approach","authors":"Yu Yan;Chengxi Zhang;Jin Wu;Chen Sun;Qingwen Meng;Chaoyi Chen;Jianqiang Wang;Guangwei Wang","doi":"10.26599/JICV.2026.9210069","DOIUrl":"https://doi.org/10.26599/JICV.2026.9210069","url":null,"abstract":"Autonomous driving on mountainous roads is challenging since numerous curves and varying road slopes make vehicle control difficult. To address this problem, we propose triple adaptive delayed deep deterministic (TAD3) policy gradient, a novel end-to-end adaptive deep reinforcement learning (DRL) system based on the twin delayed deep deterministic policy gradient (TD3) algorithm. This approach combines recurrent neural networks (RNNs) with the critic network to leverage the vehicle's historical state information and incorporates a triple-critics structure to adapt to diverse mountainous road conditions with various curvatures and gradients. Compared with previous methods, the proposed TAD3 approach achieves better vehicle performance on mountain roads and faster training speeds without extensive knowledge of vehicle dynamics. The experimental results using the TORCS simulator demonstrate that the proposed TAD3 achieves a 43%-64% smaller distance error and a 19%-74% smaller yaw angle error than five state-of-the-art baselines do in the lane-keeping task while simultaneously achieving lower laptimes in the time-minimum task and demonstrating superior generalization ability on three mountainous tracks with different designs in terms of curvature and gradient.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210069-1-9210069-14"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319348","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852504","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 : 2025-12-30DOI: 10.26599/JICV.2025.9210071
Wei Zhou;Haoye Chen;Pei Huang;Zhenliang Ma
This study proposes an integer linear program model for ride-sharing, electric, autonomous mobility on demand (RE-AMoD) system operations and develops a model predictive control (MPC) algorithm to optimize the decisions of ride matching, vehicle routing, rebalancing, and charging. The system ensures that electric autonomous vehicles provide transportation services for up to two customers to share a ride and that they can be charged automatically during the operating period. The RE-AMoD problem is formulated as a network flow optimization problem considering ride-sharing and charging control. The objective is to minimize the customers' waiting time while minimizing the system's energy consumption. An iterative MPC is developed to compute the optimal control policy for real-time control. The case study uses real-world data from San Francisco to validate the model performance by comparing benchmark models in an RE-AMoD simulation platform and investigating the impact of ride-sharing and smart charging strategies on system performance by comparing models with no ride-sharing and heuristic charging strategies. The results show that the smart charging policy is critical for realizing ride-sharing's full advantages in RE-AMoD systems. Allowing the sharing of trips significantly improves system performance in terms of reducing fleet sizes and energy consumption while improving the customer level of service.
{"title":"Integrated Operation and Charging Controls for Ride-Sharing Electric Autonomous Mobility-on-Demand Systems","authors":"Wei Zhou;Haoye Chen;Pei Huang;Zhenliang Ma","doi":"10.26599/JICV.2025.9210071","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210071","url":null,"abstract":"This study proposes an integer linear program model for ride-sharing, electric, autonomous mobility on demand (RE-AMoD) system operations and develops a model predictive control (MPC) algorithm to optimize the decisions of ride matching, vehicle routing, rebalancing, and charging. The system ensures that electric autonomous vehicles provide transportation services for up to two customers to share a ride and that they can be charged automatically during the operating period. The RE-AMoD problem is formulated as a network flow optimization problem considering ride-sharing and charging control. The objective is to minimize the customers' waiting time while minimizing the system's energy consumption. An iterative MPC is developed to compute the optimal control policy for real-time control. The case study uses real-world data from San Francisco to validate the model performance by comparing benchmark models in an RE-AMoD simulation platform and investigating the impact of ride-sharing and smart charging strategies on system performance by comparing models with no ride-sharing and heuristic charging strategies. The results show that the smart charging policy is critical for realizing ride-sharing's full advantages in RE-AMoD systems. Allowing the sharing of trips significantly improves system performance in terms of reducing fleet sizes and energy consumption while improving the customer level of service.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 4","pages":"9210071-1-9210071-13"},"PeriodicalIF":7.8,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319350","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852548","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 : 2025-09-01DOI: 10.26599/JICV.2025.9210061
Baozhen Yao;Zhihao Qi;Ziqi Liu;Minke Zhu;Shaohua Cui;Radu-Emil Precup;Raul-Cristian Roman
Environmental sustainability is a crucial issue for all human beings, and vehicle emissions significantly contribute to climate change. This has prompted many countries, including China, Norway, and Germany, to focus on electrifying transportation. This study quantifies the life cycle carbon dioxide (CO2) emissions of electric buses (EBs) in Guangzhou, China, via a life cycle analysis methodology, revealing an average life cycle emission of 1,097.07 g CO2·km−1·vehicle−1. The operation and charging stage contributes the most to the lifespan of CO2 emissions at 69.6%, driven by carbon-intensive power grid. Compared with conventional internal combustion engine buses, EBs result in significant emission reductions, but regional grid carbon intensity variations across China mean that their benefits depend on nationwide green energy adoption. By 2030, emissions are projected to decline by 15.28%, aligning with carbon peak goals. The findings emphasize that transitioning to renewable energy grids and hybrid technologies is critical for sustainable transportation.
环境可持续性对全人类来说都是一个至关重要的问题，而汽车排放对气候变化的影响很大。这促使包括中国、挪威和德国在内的许多国家将重点放在电气化交通上。本研究通过生命周期分析方法量化了中国广州电动公交车（EBs）的生命周期二氧化碳（CO2）排放量，发现其平均生命周期排放量为1,097.07 g CO2·km−1·vehicle−1。在碳密集型电网的驱动下，运行和充电阶段对二氧化碳排放寿命的贡献最大，为69.6%。与传统内燃机公交车相比，电动汽车的减排效果显著，但中国各地电网碳强度的区域差异意味着，电动汽车的效益取决于全国范围内对绿色能源的采用。到2030年，预计排放量将下降15.28%，与碳峰值目标保持一致。研究结果强调，向可再生能源电网和混合技术过渡对可持续交通至关重要。
{"title":"Evaluating the Sustainability of Electric Buses During Operation via Field Data","authors":"Baozhen Yao;Zhihao Qi;Ziqi Liu;Minke Zhu;Shaohua Cui;Radu-Emil Precup;Raul-Cristian Roman","doi":"10.26599/JICV.2025.9210061","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210061","url":null,"abstract":"Environmental sustainability is a crucial issue for all human beings, and vehicle emissions significantly contribute to climate change. This has prompted many countries, including China, Norway, and Germany, to focus on electrifying transportation. This study quantifies the life cycle carbon dioxide (CO<inf>2</inf>) emissions of electric buses (EBs) in Guangzhou, China, via a life cycle analysis methodology, revealing an average life cycle emission of 1,097.07 g CO<inf>2</inf>·km<sup>−1</sup>·vehicle<sup>−1</sup>. The operation and charging stage contributes the most to the lifespan of CO<inf>2</inf> emissions at 69.6%, driven by carbon-intensive power grid. Compared with conventional internal combustion engine buses, EBs result in significant emission reductions, but regional grid carbon intensity variations across China mean that their benefits depend on nationwide green energy adoption. By 2030, emissions are projected to decline by 15.28%, aligning with carbon peak goals. The findings emphasize that transitioning to renewable energy grids and hybrid technologies is critical for sustainable transportation.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 3","pages":"9210061-1-9210061-17"},"PeriodicalIF":7.8,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11215933","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145341074","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 intensity levels of autonomous vehicles should be thoroughly evaluated before deployment, while vehicle tests are difficult for the sake of heavy experimental resources and large numbers of cases, especially tests that include safety-critical scenarios. In this study, a new scenario generation method is proposed to accelerate the test, which is based on a multiagent reinforcement learning (MARL) framework incorporating the driving potential field (DPF). This framework is used to train some background vehicles to enable high-risk and marginal scenes, where the DPF is applied to enact the rewards of the adversarial background agents. Other background vehicles that use reasonable driving policies, which serve as naturalistic agents to increase scenario diversity, are also considered. The coexistence of naturalistic and adversarial agents enriches the experiences learned by the background cars, providing more marginal and risky scenarios for accelerating the test. The experimental results demonstrate the efficiency of the generation of high-risk and marginal scenes, with comprehensive assessments via a novel field-based dynamic risk evaluation method.
{"title":"Safety-Critical Scenario Test for Intelligent Vehicles via Hybrid Participation of Natural and Adversarial Agents","authors":"Yong Wang;Daifeng Zhang;Yanqiang Li;Liguo Shuai;Zhicheng Tang;Yuxiang Hou","doi":"10.26599/JICV.2025.9210066","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210066","url":null,"abstract":"The intensity levels of autonomous vehicles should be thoroughly evaluated before deployment, while vehicle tests are difficult for the sake of heavy experimental resources and large numbers of cases, especially tests that include safety-critical scenarios. In this study, a new scenario generation method is proposed to accelerate the test, which is based on a multiagent reinforcement learning (MARL) framework incorporating the driving potential field (DPF). This framework is used to train some background vehicles to enable high-risk and marginal scenes, where the DPF is applied to enact the rewards of the adversarial background agents. Other background vehicles that use reasonable driving policies, which serve as naturalistic agents to increase scenario diversity, are also considered. The coexistence of naturalistic and adversarial agents enriches the experiences learned by the background cars, providing more marginal and risky scenarios for accelerating the test. The experimental results demonstrate the efficiency of the generation of high-risk and marginal scenes, with comprehensive assessments via a novel field-based dynamic risk evaluation method.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 3","pages":"9210066-1-9210066-13"},"PeriodicalIF":7.8,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11215927","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145339719","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}
With increasing autonomous vehicle (AV) penetration, understanding the factors contributing to AV crashes is crucial for addressing ongoing road safety challenges. This study aims to reveal the effects of vehicle characteristics, road conditions, environmental factors, and precrash movements on the occurrence of head-on, rear-end, and side-impact crashes. In particular, factors associated with various types of Level 2 and Level 4 AV crashes were analyzed. Our data, obtained from the California Department of Motor Vehicles and the National Highway Traffic Safety Administration, spans from October 2014 to March 2024. Association rule mining techniques identify the significant patterns and interdependencies among the factors contributing to AV crashes. The findings demonstrate that the factors influencing crash types include weather, roadway surface, lighting, and vehicle precrash movements. For example, head-on crashes frequently occur at intersections under poor lighting conditions, whereas rear-end crashes occur more frequently on high-speed highways, particularly because of unexpected braking by the vehicle ahead. Side-impact crashes commonly result from merging maneuvers, especially under adverse weather and lighting conditions. These findings provide new insights into the causal mechanisms behind different types of AV crashes and underscore the need to strengthen traffic management, enhance AV sensing and decision-making capabilities, and implement targeted safety measures in high-risk areas.
{"title":"What Patterns Contribute to Autonomous Vehicle Crashes? A Study of Level 2 and 4 Automation via Association Rule Analysis","authors":"Hongliang Ding;Sicong Wang;Yang Cao;Xiaowen Fu;Hanlong Fu;Quan Yuan;Tiantian Chen","doi":"10.26599/JICV.2025.9210065","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210065","url":null,"abstract":"With increasing autonomous vehicle (AV) penetration, understanding the factors contributing to AV crashes is crucial for addressing ongoing road safety challenges. This study aims to reveal the effects of vehicle characteristics, road conditions, environmental factors, and precrash movements on the occurrence of head-on, rear-end, and side-impact crashes. In particular, factors associated with various types of Level 2 and Level 4 AV crashes were analyzed. Our data, obtained from the California Department of Motor Vehicles and the National Highway Traffic Safety Administration, spans from October 2014 to March 2024. Association rule mining techniques identify the significant patterns and interdependencies among the factors contributing to AV crashes. The findings demonstrate that the factors influencing crash types include weather, roadway surface, lighting, and vehicle precrash movements. For example, head-on crashes frequently occur at intersections under poor lighting conditions, whereas rear-end crashes occur more frequently on high-speed highways, particularly because of unexpected braking by the vehicle ahead. Side-impact crashes commonly result from merging maneuvers, especially under adverse weather and lighting conditions. These findings provide new insights into the causal mechanisms behind different types of AV crashes and underscore the need to strengthen traffic management, enhance AV sensing and decision-making capabilities, and implement targeted safety measures in high-risk areas.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 3","pages":"9210065-1-9210065-16"},"PeriodicalIF":7.8,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11215932","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145341067","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 routing problem for intelligent and connected vehicles has garnered significant attention because of its profound theoretical implications and wide-ranging practical applications. Despite advancements, existing learning-based methods often rely on training one single policy, which inadequately explores the solution space and leads to suboptimal performance. To address this limitation, we propose a diversified tour-driven deep reinforcement learning (DT-DRL) approach for solving vehicle routing problems (VRPs) across various scales. Our approach builds on the encoder-decoder paradigm, with the encoder utilizing a multihead attention mechanism to derive informative node embeddings and a gate aggregation block to enhance state representation. During decoding, dynamic-aware context embedding is designed to capture real-time state transitions and graph variations, thereby offering comprehensive and timely information for decision-making. To promote solution diversity and expand the search space, multiple decoders with independent parameters are employed, coupled with a Kullback-Leibler divergence-based cross-entropy loss that regularizes the generation of diversified candidate tours. We validate the proposed DT-DRL through extensive experimentation on two representative routing problems for intelligent connected vehicles, namely, the traveling salesman problem (TSP) and the capacitated VRP (CVRP). The results demonstrate that DT-DRL consistently outperforms many heuristic and DRL-based methods, achieving up to a 7.54% improvement in the optimality gap, thereby establishing its effectiveness and robustness in tackling complex routing challenges for intelligent and connected vehicles.
{"title":"A Diversified Tour-Driven Deep Reinforcement Learning Approach to Routing for Intelligent and Connected Vehicles","authors":"Dapeng Yan;Qingshu Guan;Bei Ou;Bowen Yan;Hui Cao;Badong Chen","doi":"10.26599/JICV.2025.9210064","DOIUrl":"https://doi.org/10.26599/JICV.2025.9210064","url":null,"abstract":"The routing problem for intelligent and connected vehicles has garnered significant attention because of its profound theoretical implications and wide-ranging practical applications. Despite advancements, existing learning-based methods often rely on training one single policy, which inadequately explores the solution space and leads to suboptimal performance. To address this limitation, we propose a diversified tour-driven deep reinforcement learning (DT-DRL) approach for solving vehicle routing problems (VRPs) across various scales. Our approach builds on the encoder-decoder paradigm, with the encoder utilizing a multihead attention mechanism to derive informative node embeddings and a gate aggregation block to enhance state representation. During decoding, dynamic-aware context embedding is designed to capture real-time state transitions and graph variations, thereby offering comprehensive and timely information for decision-making. To promote solution diversity and expand the search space, multiple decoders with independent parameters are employed, coupled with a Kullback-Leibler divergence-based cross-entropy loss that regularizes the generation of diversified candidate tours. We validate the proposed DT-DRL through extensive experimentation on two representative routing problems for intelligent connected vehicles, namely, the traveling salesman problem (TSP) and the capacitated VRP (CVRP). The results demonstrate that DT-DRL consistently outperforms many heuristic and DRL-based methods, achieving up to a 7.54% improvement in the optimality gap, thereby establishing its effectiveness and robustness in tackling complex routing challenges for intelligent and connected vehicles.","PeriodicalId":100793,"journal":{"name":"Journal of Intelligent and Connected Vehicles","volume":"8 3","pages":"9210064-1-9210064-11"},"PeriodicalIF":7.8,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11215934","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145339697","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：