Multimodal Transportation最新文献_第3页

Catboost - SHapley Additive exPlanations (SHAP) car following model: Explaining model features in mixed traffic conditions Catboost - SHapley加性解释（SHAP）汽车跟随模型：解释混合交通条件下的模型特征

Multimodal Transportation

Pub Date : 2025-11-28 DOI: 10.1016/j.multra.2025.100284

Tianyang Cui , D. J. Wu , Zejiang Wang

Car-following behavior is critical for traffic management, road design, and the development of advanced driver assistance systems (ADAS) and autonomous vehicles (AVs). Traditional theory-based car-following models are widely used in traffic simulations but rely on simplified assumptions, limiting their ability to capture the complexity of real-world driving. In contrast, machine learning (ML) models can leverage large datasets to uncover complex driving behaviors. However, a major limitation of ML models is their lack of interpretability. Moreover, the rise of AVs has introduced mixed-traffic environments where AVs and human-driven vehicles share the road. Understanding different interaction scenarios—such as AVs following human drivers (AH), human drivers following AVs (HA), and human drivers following other humans (HH)—is essential for accurate modeling and safe AV deployment. To address these challenges, we propose a car-following modeling framework that integrates the CatBoost algorithm with SHapley Additive exPlanations (SHAP). CatBoost handles both numerical and categorical data, enabling the development of scenario-specific models (AH, HA, HH) and a unified car-following model incorporating scenario type as a feature. SHAP enhances interpretability by quantifying the contribution of each model feature, e.g., speed and inter-vehicle distance, across scenarios. We apply this framework to the Lyft Level-5 dataset to analyze feature importance and evaluate how scenario type moderates driving behavior. The insights derived from our analysis support the design of more adaptive AV control strategies and inform transportation policies for the safe integration of AVs into modern traffic systems.

车辆跟随行为对于交通管理、道路设计以及先进驾驶辅助系统（ADAS）和自动驾驶汽车（AVs）的开发至关重要。传统的基于理论的汽车跟随模型广泛应用于交通模拟，但依赖于简化的假设，限制了它们捕捉真实驾驶复杂性的能力。相比之下，机器学习（ML）模型可以利用大型数据集来揭示复杂的驾驶行为。然而，ML模型的一个主要限制是它们缺乏可解释性。此外，自动驾驶汽车的兴起引入了混合交通环境，自动驾驶汽车和人类驾驶的车辆共享道路。了解不同的交互场景——例如自动驾驶汽车跟随人类驾驶员（AH），人类驾驶员跟随自动驾驶汽车（HA），以及人类驾驶员跟随其他人（HH）——对于准确建模和安全部署自动驾驶汽车至关重要。为了解决这些挑战，我们提出了一个将CatBoost算法与SHapley加性解释（SHAP）集成在一起的汽车跟随建模框架。CatBoost可以处理数值和分类数据，从而能够开发特定于场景的模型（AH， HA， HH）和将场景类型作为特征的统一汽车跟随模型。SHAP通过量化每个模型特征的贡献来增强可解释性，例如跨场景的速度和车辆间距离。我们将此框架应用于Lyft Level-5数据集，以分析特征的重要性，并评估场景类型如何调节驾驶行为。从我们的分析中得出的见解支持设计更具适应性的自动驾驶控制策略，并为将自动驾驶汽车安全集成到现代交通系统中的交通政策提供信息。

{"title":"Catboost - SHapley Additive exPlanations (SHAP) car following model: Explaining model features in mixed traffic conditions","authors":"Tianyang Cui , D. J. Wu , Zejiang Wang","doi":"10.1016/j.multra.2025.100284","DOIUrl":"10.1016/j.multra.2025.100284","url":null,"abstract":"<div><div>Car-following behavior is critical for traffic management, road design, and the development of advanced driver assistance systems (ADAS) and autonomous vehicles (AVs). Traditional theory-based car-following models are widely used in traffic simulations but rely on simplified assumptions, limiting their ability to capture the complexity of real-world driving. In contrast, machine learning (ML) models can leverage large datasets to uncover complex driving behaviors. However, a major limitation of ML models is their lack of interpretability. Moreover, the rise of AVs has introduced mixed-traffic environments where AVs and human-driven vehicles share the road. Understanding different interaction scenarios—such as AVs following human drivers (AH), human drivers following AVs (HA), and human drivers following other humans (HH)—is essential for accurate modeling and safe AV deployment. To address these challenges, we propose a car-following modeling framework that integrates the CatBoost algorithm with SHapley Additive exPlanations (SHAP). CatBoost handles both numerical and categorical data, enabling the development of scenario-specific models (AH, HA, HH) and a unified car-following model incorporating scenario type as a feature. SHAP enhances interpretability by quantifying the contribution of each model feature, e.g., speed and inter-vehicle distance, across scenarios. We apply this framework to the Lyft Level-5 dataset to analyze feature importance and evaluate how scenario type moderates driving behavior. The insights derived from our analysis support the design of more adaptive AV control strategies and inform transportation policies for the safe integration of AVs into modern traffic systems.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 2","pages":"Article 100284"},"PeriodicalIF":0.0,"publicationDate":"2025-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145693538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A bi-objective slot allocation model under airport capacity and resource utilization 考虑机场容量和资源利用的双目标机位分配模型

Multimodal Transportation

Pub Date : 2025-11-15 DOI: 10.1016/j.multra.2025.100273

Priyadarshan Loshan, Loshaka Perera

Most airports operate below their declared capacity, yet expansion through costly infrastructure development remains the primary strategy for meeting the rising demand for flights. Inefficient slot allocation, underutilised airside resources, and a lack of detailed demand–capacity analysis hinder performance, leading to rejected slot requests that may cost the industry billions of dollars annually. This research proposes a practical alternative: optimizing existing capacity before pursuing expansion using a bi-objective mathematical model. This model simultaneously maximizes the utilisation of runway, apron, and terminal gate capacity through revised slot scheduling, and incorporates real-time operational constraints to minimise delay propagation while maintaining separation minima. The model was validated using real data from Bandaranaike International Airport-Colombo (BIA). The proposed linear programming model demonstrated increased average airside resource utilisation on peak days, from 44.9% to 62.7%, while ensuring that the current schedule peak traffic intensities are maintained. Through delay optimization, the proposed schedule is capable of reducing congestion and cumulative delays compared to the non-optimized schedule, mainly when delays are propagated due to uncertainties. With an average delay reduction of 140.80 min per scenario, the model's validity was confirmed, providing strong evidence of its robustness and reliability. These results demonstrate the potential of optimized slot allocation as a decision-support tool, enabling fairer access for new entrants, reducing delays, and enhancing efficiency across existing operations.

大多数机场的运营低于其宣布的运力，但通过昂贵的基础设施建设进行扩张仍然是满足不断增长的航班需求的主要策略。低效的机位分配、未充分利用的空侧资源以及缺乏详细的需求-容量分析都阻碍了性能，导致机位请求被拒绝，这可能导致该行业每年损失数十亿美元。本研究提出了一种实用的替代方案：利用双目标数学模型优化现有产能，然后再追求扩张。该模型通过修正的机位调度，同时最大限度地利用跑道、停机坪和航站楼登机口的容量，并结合实时操作约束，以最大限度地减少延误传播，同时保持最小的间隔。该模型使用班达拉奈克-科伦坡国际机场（BIA）的真实数据进行了验证。建议的线性规划模型显示，在高峰日，平均空侧资源利用率从44.9%提高到62.7%，同时确保维持目前的时间表高峰交通强度。通过延迟优化，与未优化的调度相比，提出的调度能够减少拥塞和累积延迟，主要是当延迟由于不确定性而传播时。每个场景平均延迟减少140.80 min，验证了模型的有效性，有力地证明了模型的鲁棒性和可靠性。这些结果证明了优化时段分配作为决策支持工具的潜力，可以为新进入者提供更公平的准入，减少延误，并提高现有业务的效率。

{"title":"A bi-objective slot allocation model under airport capacity and resource utilization","authors":"Priyadarshan Loshan, Loshaka Perera","doi":"10.1016/j.multra.2025.100273","DOIUrl":"10.1016/j.multra.2025.100273","url":null,"abstract":"<div><div>Most airports operate below their declared capacity, yet expansion through costly infrastructure development remains the primary strategy for meeting the rising demand for flights. Inefficient slot allocation, underutilised airside resources, and a lack of detailed demand–capacity analysis hinder performance, leading to rejected slot requests that may cost the industry billions of dollars annually. This research proposes a practical alternative: optimizing existing capacity before pursuing expansion using a bi-objective mathematical model. This model simultaneously maximizes the utilisation of runway, apron, and terminal gate capacity through revised slot scheduling, and incorporates real-time operational constraints to minimise delay propagation while maintaining separation minima. The model was validated using real data from Bandaranaike International Airport-Colombo (BIA). The proposed linear programming model demonstrated increased average airside resource utilisation on peak days, from 44.9% to 62.7%, while ensuring that the current schedule peak traffic intensities are maintained. Through delay optimization, the proposed schedule is capable of reducing congestion and cumulative delays compared to the non-optimized schedule, mainly when delays are propagated due to uncertainties. With an average delay reduction of 140.80 min per scenario, the model's validity was confirmed, providing strong evidence of its robustness and reliability. These results demonstrate the potential of optimized slot allocation as a decision-support tool, enabling fairer access for new entrants, reducing delays, and enhancing efficiency across existing operations.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 2","pages":"Article 100273"},"PeriodicalIF":0.0,"publicationDate":"2025-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145623482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Regulating jaywalking behaviour in adaptive traffic signal control using a novel deep reinforcement learning approach 基于深度强化学习的自适应交通信号控制中乱穿马路行为的调节

Multimodal Transportation

Pub Date : 2025-11-11 DOI: 10.1016/j.multra.2025.100271

Lok Sang Chan, Xiaocai Zhang, Neema Nassir, Majid Sarvi

This paper presents a deep reinforcement learning based adaptive traffic signal control framework that explicitly models jaywalking at urban intersections. We integrate a behaviourally grounded Jaywalking Decision model, which endogenises red light violations through waiting time and dynamic gap acceptance, with a Branching Double Deep Q-Network and a comprehensive hybrid action space that controls both phase selection and subphase timing. A multiobjective reward balances delay and jaywalking related safety risk, enabling the controller to respond to non-compliance as it emerges. The framework is evaluated in a multimodal microsimulation of a real intersection in Melbourne across four naturalistic demand scenarios and against both actuated control and established reinforcement learning baselines. Results show that the proposed approach reduces observed pedestrian delay and jaywalking relative to actuated control while achieving a more balanced safety and efficiency profile than single objective or alternative learning architectures. The analysis highlights context-dependent trade-offs that are relevant for policy, since the controller can adapt timing to mitigate non-compliance without assuming full pedestrian obedience. The contributions are threefold: a realistic jaywalking model linked to observable states, a high-granularity action space for multimodal control, and an integrated learning framework that jointly manages delay and safety risk. The proposed framework not only facilitates a more equitable traffic operation system but also offers the first scalable approach to managing risky behaviours in urban traffic environments.

本文提出了一种基于深度强化学习的自适应交通信号控制框架，该框架明确地模拟了城市十字路口的乱穿马路行为。我们整合了一个基于行为的乱穿马路决策模型，该模型通过等待时间和动态间隙接受来内化红灯违规行为，并集成了一个分支双深度q网络和一个控制相位选择和子相位定时的综合混合动作空间。多目标奖励平衡了延迟和乱穿马路相关的安全风险，使控制器能够在出现违规行为时做出反应。该框架在墨尔本一个真实十字路口的多模态微模拟中进行了评估，涉及四个自然需求场景，并针对驱动控制和已建立的强化学习基线。结果表明，与驱动控制相比，该方法减少了观察到的行人延迟和乱穿马路，同时比单一目标或替代学习架构实现了更平衡的安全性和效率。该分析强调了与策略相关的上下文相关的权衡，因为控制器可以调整时间以减轻不遵守规定的情况，而无需假设行人完全服从。贡献有三个方面：与可观察状态相关联的逼真的乱穿马路模型，用于多模式控制的高粒度行动空间，以及联合管理延迟和安全风险的集成学习框架。拟议的框架不仅有助于建立一个更公平的交通运营系统，而且还提供了第一个可扩展的方法来管理城市交通环境中的危险行为。

{"title":"Regulating jaywalking behaviour in adaptive traffic signal control using a novel deep reinforcement learning approach","authors":"Lok Sang Chan, Xiaocai Zhang, Neema Nassir, Majid Sarvi","doi":"10.1016/j.multra.2025.100271","DOIUrl":"10.1016/j.multra.2025.100271","url":null,"abstract":"<div><div>This paper presents a deep reinforcement learning based adaptive traffic signal control framework that explicitly models jaywalking at urban intersections. We integrate a behaviourally grounded Jaywalking Decision model, which endogenises red light violations through waiting time and dynamic gap acceptance, with a Branching Double Deep Q-Network and a comprehensive hybrid action space that controls both phase selection and subphase timing. A multiobjective reward balances delay and jaywalking related safety risk, enabling the controller to respond to non-compliance as it emerges. The framework is evaluated in a multimodal microsimulation of a real intersection in Melbourne across four naturalistic demand scenarios and against both actuated control and established reinforcement learning baselines. Results show that the proposed approach reduces observed pedestrian delay and jaywalking relative to actuated control while achieving a more balanced safety and efficiency profile than single objective or alternative learning architectures. The analysis highlights context-dependent trade-offs that are relevant for policy, since the controller can adapt timing to mitigate non-compliance without assuming full pedestrian obedience. The contributions are threefold: a realistic jaywalking model linked to observable states, a high-granularity action space for multimodal control, and an integrated learning framework that jointly manages delay and safety risk. The proposed framework not only facilitates a more equitable traffic operation system but also offers the first scalable approach to managing risky behaviours in urban traffic environments.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 2","pages":"Article 100271"},"PeriodicalIF":0.0,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579807","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Towards autonomy: A comprehensive technical and ethical review of automated vehicle safety 走向自主：对自动驾驶汽车安全的全面技术和伦理审查

Multimodal Transportation

Pub Date : 2025-11-10 DOI: 10.1016/j.multra.2025.100272

Morteza Soleimani, Ayse Aysu Sari

This review paper explores the dynamic landscape of Automated Vehicles (AVs), with safety as a top priority. It discusses various dimensions, encompassing the pivotal role of safety in AV development, regulatory frameworks guiding their operation, intricacies of software architecture designed to uphold safety standards, and societal perceptions. The review scrutinizes diverse methodologies employed in ensuring vehicle safety, advocating for a holistic integration of safety and security. It calls for collaborative efforts to enhance the safety and security of AVs even more. Addressing current gaps in safety regulations, the paper outlines new safety approaches designed specifically for AVs. Furthermore, it discusses expert-recommended strategies for strong software engineering in AV systems, aimed at improving reliability. The review also considers the profound implications of AV integration in society and legal frameworks. By bringing together different viewpoints, it aims to build a clear understanding of AV safety while identifying new challenges in this fast-changing field of engineering.

这篇综述探讨了自动驾驶汽车（AVs）的动态景观，安全是重中之重。它讨论了各个方面，包括安全在自动驾驶开发中的关键作用，指导其操作的监管框架，旨在维护安全标准的软件架构的复杂性，以及社会观念。该检讨检讨了确保车辆安全所采用的各种方法，提倡安全与保安的整体整合。它呼吁各方共同努力，进一步提高自动驾驶汽车的安全性。为了解决当前安全法规的空白，该论文概述了专门为自动驾驶汽车设计的新安全方法。此外，它还讨论了专家推荐的AV系统强大软件工程策略，旨在提高可靠性。该审查还考虑了AV整合在社会和法律框架中的深远影响。通过汇集不同的观点，旨在建立对自动驾驶安全的清晰认识，同时在这个快速变化的工程领域确定新的挑战。

引用次数: 0

A method for determining pickup and delivery locations of intercity customized bus based on passenger demand and POIs 基于乘客需求和poi的城际定制巴士接送地点确定方法

Multimodal Transportation

Pub Date : 2025-10-24 DOI: 10.1016/j.multra.2025.100270

Yueying Huo , Huijuan Zhou , Feng Hao , Man Zhang , Yachao Liu

Intercity customized bus is a new mode of road passenger transport that relies on the internet platform to obtain passengers' reservation travel demands and provide passengers with “door-to-door” transport service between cities. The determining of Pickup and Delivery locations is essential for its operation, as it provides the possibility of “door-to-door” direct transport service. Existing methods for determining Pickup and Delivery locations mainly focus on clustering passenger demand data, which will lead to the problem of passengers and drivers having difficulty in quickly finding sites in the road network. Therefore, this study aims to propose a new method for determining Pickup and Delivery locations both considering passenger demand data and POIs. Based on the passenger reservation data and AutoNavi Map API, suitable POI categories are selected to derive the actual walking distances and routes between passengers and different POIs. Through two rounds of screening, The POIs with the wider service coverage and the smallest actual walking distance for passengers was selected as the sites. The results show that by utilizing the new method for determining Pickup and Delivery locations, we identified the locations of suitable sites and controlled the actual walking distance of passengers within 500 m in the road network, which will provide convenience to both drivers and passengers. This study will provide a reference basis for optimizing the site setting in intercity customized bus.

城际定制客车是依托互联网平台获取乘客预定出行需求，为乘客提供城际“门到门”运输服务的一种新型道路客运模式。取货地点的确定对其运营至关重要，因为它提供了“门到门”直接运输服务的可能性。现有的取车和送车地点的确定方法主要集中在聚集乘客需求数据上，这将导致乘客和司机难以在路网中快速找到地点的问题。因此，本研究旨在提出一种考虑乘客需求数据和poi的新方法来确定取货和交货地点。根据乘客预订数据和高德地图API，选择合适的POI类别，得出乘客与不同POI之间的实际步行距离和路线。经过两轮筛选，我们选择服务范围较广、乘客实际步行距离最短的机场作为选址。结果表明：利用新的取货地点确定方法，我们确定了合适的地点位置，并将乘客在路网中的实际步行距离控制在500 m以内，这将为司机和乘客提供便利。本研究将为城际定制巴士站点设置的优化提供参考依据。

{"title":"A method for determining pickup and delivery locations of intercity customized bus based on passenger demand and POIs","authors":"Yueying Huo , Huijuan Zhou , Feng Hao , Man Zhang , Yachao Liu","doi":"10.1016/j.multra.2025.100270","DOIUrl":"10.1016/j.multra.2025.100270","url":null,"abstract":"<div><div>Intercity customized bus is a new mode of road passenger transport that relies on the internet platform to obtain passengers' reservation travel demands and provide passengers with “door-to-door” transport service between cities. The determining of Pickup and Delivery locations is essential for its operation, as it provides the possibility of “door-to-door” direct transport service. Existing methods for determining Pickup and Delivery locations mainly focus on clustering passenger demand data, which will lead to the problem of passengers and drivers having difficulty in quickly finding sites in the road network. Therefore, this study aims to propose a new method for determining Pickup and Delivery locations both considering passenger demand data and POIs. Based on the passenger reservation data and AutoNavi Map API, suitable POI categories are selected to derive the actual walking distances and routes between passengers and different POIs. Through two rounds of screening, The POIs with the wider service coverage and the smallest actual walking distance for passengers was selected as the sites. The results show that by utilizing the new method for determining Pickup and Delivery locations, we identified the locations of suitable sites and controlled the actual walking distance of passengers within 500 m in the road network, which will provide convenience to both drivers and passengers. This study will provide a reference basis for optimizing the site setting in intercity customized bus.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 2","pages":"Article 100270"},"PeriodicalIF":0.0,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145529075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic workforce allocation for train-based crowd-shipping: Multi-period optimization incorporating transit discomfort costs 基于列车的人群运输动态劳动力分配：考虑运输不适成本的多周期优化

Multimodal Transportation

Pub Date : 2025-10-16 DOI: 10.1016/j.multra.2025.100269

Qiuhong Huang, Shinya Hanaoka

Last-mile parcel delivery faces challenges such as rising costs, labor shortages, and environmental concerns, especially in urban areas. This paper tackles these issues with a new rolling horizon framework that uses public train networks. We create a dynamic model that assigns delivery tasks to crowd-shippers (commuters) and part-time workers based on real-time data, considering train occupancy and commuter discomfort costs. Our results show that strategic delays in parcel delivery can offer economic benefits, with a delayed end-of-day approach saving 6.3% to 9.1% compared to the faster mixed-workforce approach we studied, which involved both part-timers and crowd-shippers. Additionally, we find that crowdsourced delivery is mainly cost-effective during off-peak transit hours, while part-time staff are more economical during peak times. These findings support adopting a flexible, hybrid workforce strategy that adjusts to changing transit conditions over time.

最后一英里的包裹递送面临着成本上升、劳动力短缺和环境问题等挑战，尤其是在城市地区。本文用一个新的滚动地平线框架解决了这些问题，该框架使用公共列车网络。我们创建了一个动态模型，根据实时数据，考虑到火车占用率和通勤者的不适成本，将送货任务分配给人群运送者（通勤者）和兼职工人。我们的研究结果表明，战略性延迟包裹递送可以提供经济效益，与我们研究的更快的混合劳动力方法相比，延迟的一天结束方法节省了6.3%至9.1%，混合劳动力方法涉及兼职人员和人群快递员。此外，我们发现众包配送主要在非高峰运输时段具有成本效益，而兼职人员在高峰时段更为经济。这些发现支持采用灵活的混合劳动力战略，以适应不断变化的交通条件。

引用次数: 0

DrivNet: Adaptive driving behavior prediction for improving the safety of multimodal transportation in autonomous vehicle systems DrivNet：自适应驾驶行为预测，用于提高自动驾驶车辆系统中多式联运的安全性

Multimodal Transportation

Pub Date : 2025-09-18 DOI: 10.1016/j.multra.2025.100266

J Robert Theivadas, Dr Suresh Ponnan

The safety and efficiency of autonomous vehicles depend on their ability to react swiftly and appropriately to dynamic driving environments. This paper presents DrivNet (Adaptive Driving Network), an adaptive algorithm, DrivNet, aimed at enhancing the reaction time and decision-making capabilities of autonomous vehicles under varying traffic conditions and times of day. DrivNet integrates vehicle parameters and driving behavior with physiological features to improve forward movement within situational awareness environments. The algorithm is validated using a real-time driving dataset collected from expert drivers, capturing both vehicle and driving behavior data under heavy and normal traffic conditions, as well as day and night scenarios. Distinct driving behavior patterns are generated for three key situational awareness conditions: accelerating on clear roads, navigating through critical situations, and maintaining fuel efficiency. These patterns serve as the basis for adapting vehicle control decisions. To validate the effectiveness of these driving behavior patterns, a Recurrent Neural Network (RNN) architecture is employed, enabling the detection and classification of psychological features such as mental workload, stress, and fatigue. The proposed DrivNet algorithm offers valuable insights into distinguishing safe from unsafe driving modes, thereby supporting an intelligent control mechanism that enhances the overall safety of autonomous transportation systems. The results demonstrate the potential of DrivNet to improve autonomous vehicle performance, contributing to the future of safe and efficient self-driving technologies.

自动驾驶汽车的安全性和效率取决于它们对动态驾驶环境做出快速、适当反应的能力。本文提出了自适应驾驶网络（Adaptive Driving Network），一种旨在提高自动驾驶汽车在不同交通条件和时间下的反应时间和决策能力的自适应算法。DrivNet将车辆参数和驾驶行为与生理特征相结合，以改善在态势感知环境下的前进运动。该算法使用从专业驾驶员那里收集的实时驾驶数据集进行验证，该数据集捕获了在繁忙和正常交通条件下以及白天和夜间场景下的车辆和驾驶行为数据。在三种关键的态势感知条件下，会产生不同的驾驶行为模式：在空旷的道路上加速，在紧急情况下导航，以及保持燃油效率。这些模式作为调整车辆控制决策的基础。为了验证这些驾驶行为模式的有效性，采用了循环神经网络（RNN）架构，可以检测和分类心理特征，如心理工作量、压力和疲劳。提出的DrivNet算法为区分安全与不安全驾驶模式提供了有价值的见解，从而支持智能控制机制，提高自动运输系统的整体安全性。测试结果证明了drivenet在提高自动驾驶汽车性能方面的潜力，为未来安全高效的自动驾驶技术做出了贡献。

{"title":"DrivNet: Adaptive driving behavior prediction for improving the safety of multimodal transportation in autonomous vehicle systems","authors":"J Robert Theivadas, Dr Suresh Ponnan","doi":"10.1016/j.multra.2025.100266","DOIUrl":"10.1016/j.multra.2025.100266","url":null,"abstract":"<div><div>The safety and efficiency of autonomous vehicles depend on their ability to react swiftly and appropriately to dynamic driving environments. This paper presents DrivNet (Adaptive Driving Network), an adaptive algorithm, DrivNet, aimed at enhancing the reaction time and decision-making capabilities of autonomous vehicles under varying traffic conditions and times of day. DrivNet integrates vehicle parameters and driving behavior with physiological features to improve forward movement within situational awareness environments. The algorithm is validated using a real-time driving dataset collected from expert drivers, capturing both vehicle and driving behavior data under heavy and normal traffic conditions, as well as day and night scenarios. Distinct driving behavior patterns are generated for three key situational awareness conditions: accelerating on clear roads, navigating through critical situations, and maintaining fuel efficiency. These patterns serve as the basis for adapting vehicle control decisions. To validate the effectiveness of these driving behavior patterns, a Recurrent Neural Network (RNN) architecture is employed, enabling the detection and classification of psychological features such as mental workload, stress, and fatigue. The proposed DrivNet algorithm offers valuable insights into distinguishing safe from unsafe driving modes, thereby supporting an intelligent control mechanism that enhances the overall safety of autonomous transportation systems. The results demonstrate the potential of DrivNet to improve autonomous vehicle performance, contributing to the future of safe and efficient self-driving technologies.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 2","pages":"Article 100266"},"PeriodicalIF":0.0,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145420051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Defense against multi-path attacks on EV charging networks: A dependency chain analysis and DQN-BASED SOLUTION 电动汽车充电网络多路径攻击防御：依赖链分析及基于dqn的解决方案

Multimodal Transportation

Pub Date : 2025-09-03 DOI: 10.1016/j.multra.2025.100254

Xin Chen , Jiaxin Peng , Luanjuan Jiang , Jixiang Cheng , Shouyu Wu

As electric vehicle (EV) charging infrastructures evolve to become more intelligent, integrated, and interconnected, they encounter escalating cybersecurity threats. Existing studies largely emphasize single-component vulnerabilities while overlooking systemic risks arising from multi-path dependency chains. This paper conceptualizes electric vehicle charging networks as a cyber-physical system (CPS) integrated with power, communication, and control layers. To tackle the issue of multi-path attack propagation within the EV charging network, we present a risk assessment method combined with graph-based system modeling with Deep Q-Networks (DQN). The attacker’s behavior is modeled as a Markov decision process, utilizing DQN to learn optimal attack paths based on cumulative rewards. This method identifies the most vulnerable components and critical propagation pathways, facilitating the development of optimized defense strategies for the deployment of constrained security resources. Comparative experiments indicate that the proposed DQN-based defense strategy outperforms random and traditional dependency-based allocations, leading to diminished cumulative attacker rewards and enhanced network resilience through more efficient resource utilization. These findings can offer practical insights for strengthening the robustness of smart grid ecosystems against multi-stage cyberattacks.

随着电动汽车（EV）充电基础设施向智能化、集成化和互联化发展，它们面临的网络安全威胁也在不断升级。现有的研究大多强调单一组件的漏洞，而忽略了多路径依赖链带来的系统性风险。本文将电动汽车充电网络概念化为一个集成了电源、通信和控制层的网络物理系统（CPS）。为了解决电动汽车充电网络中的多路径攻击传播问题，提出了一种基于深度q网络（Deep Q-Networks， DQN）的基于图的系统建模风险评估方法。将攻击者的行为建模为马尔可夫决策过程，利用DQN学习基于累积奖励的最优攻击路径。该方法识别出最易受攻击的组件和关键传播路径，促进了约束安全资源部署的优化防御策略的发展。对比实验表明，提出的基于dqn的防御策略优于随机分配和传统的基于依赖分配，通过更有效的资源利用，减少了累积攻击者奖励，增强了网络弹性。这些发现可以为加强智能电网生态系统抵御多阶段网络攻击的稳健性提供实用的见解。

{"title":"Defense against multi-path attacks on EV charging networks: A dependency chain analysis and DQN-BASED SOLUTION","authors":"Xin Chen , Jiaxin Peng , Luanjuan Jiang , Jixiang Cheng , Shouyu Wu","doi":"10.1016/j.multra.2025.100254","DOIUrl":"10.1016/j.multra.2025.100254","url":null,"abstract":"<div><div>As electric vehicle (EV) charging infrastructures evolve to become more intelligent, integrated, and interconnected, they encounter escalating cybersecurity threats. Existing studies largely emphasize single-component vulnerabilities while overlooking systemic risks arising from multi-path dependency chains. This paper conceptualizes electric vehicle charging networks as a cyber-physical system (CPS) integrated with power, communication, and control layers. To tackle the issue of multi-path attack propagation within the EV charging network, we present a risk assessment method combined with graph-based system modeling with Deep Q-Networks (DQN). The attacker’s behavior is modeled as a Markov decision process, utilizing DQN to learn optimal attack paths based on cumulative rewards. This method identifies the most vulnerable components and critical propagation pathways, facilitating the development of optimized defense strategies for the deployment of constrained security resources. Comparative experiments indicate that the proposed DQN-based defense strategy outperforms random and traditional dependency-based allocations, leading to diminished cumulative attacker rewards and enhanced network resilience through more efficient resource utilization. These findings can offer practical insights for strengthening the robustness of smart grid ecosystems against multi-stage cyberattacks.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 1","pages":"Article 100254"},"PeriodicalIF":0.0,"publicationDate":"2025-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145420457","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Simulation-based optimization for transportation system analysis: State-of-the-art research and future endeavors 基于仿真的交通系统分析优化：最新研究和未来努力

Multimodal Transportation

Pub Date : 2025-07-18 DOI: 10.1016/j.multra.2025.100253

Ziyuan Gu , Bowei Ru , Yifan Li , Wei Ma , Hai L. Vu , Qixiu Cheng , Yuan Kuang

引用次数: 0

Autonomous vehicle adoption behavior and safety concern: A study of public perception 自动驾驶汽车使用行为与安全关注：公众认知研究

Multimodal Transportation

Pub Date : 2025-07-05 DOI: 10.1016/j.multra.2025.100252

Fatemeh Nazari , Mohamadhossein Noruzoliaee , Abolfazl (Kouros) Mohammadian

Realizing the economic and societal benefits of autonomous vehicles (AVs) hinges on widespread public acceptance. However, existing research offers limited insights into two key behavioral factors shaping AV acceptance, namely, perceived AV safety concern and travel behavior, the latter reflecting how heterogenous mobility patterns influence the AV acceptance. These factors are often treated as exogenous, limiting insight into their true behavioral interdependencies with AV acceptance and their distinct behavioral roots. This study addresses these gaps by introducing a recursive trivariate econometric model that jointly estimates AV acceptance, perceived safety concern, and current travel behavior (proxied by annual vehicle-miles traveled or VMT). The recursive structure accounts for structural endogeneity, enabling the unbiased estimation of how safety concern and travel behavior influence AV acceptance, while treating both as endogenous constructs shaped by their own determinants. To further enhance behavioral realism, the model incorporates latent psychological constructs using structural equation modeling. Empirical results from a California stated preference dataset highlight that safety concern and latent vehicle cost consciousness are the two dominant deterrents to AV acceptance, suggesting that policies such as trust-building campaigns and financial incentives can stimulate AV acceptance. Despite showing less safety concern, high-VMT individuals exhibit lower AV acceptance, suggesting potential habitual inertia in ceding driving control and challenging conjectures that users embrace in-vehicle saving and that AVs promote urban sprawl. Shared mobility enthusiasm and latent vehicle performance preference alleviate AV safety concern. Gender and racial gaps persist, with women expressing greater safety concerns and Asians exhibiting higher AV acceptance.

实现自动驾驶汽车的经济和社会效益取决于公众的广泛接受。然而，现有研究对影响自动驾驶汽车接受度的两个关键行为因素，即感知自动驾驶汽车安全关注和出行行为的见解有限，后者反映了异质性出行模式如何影响自动驾驶汽车接受度。这些因素通常被视为外源性因素，限制了对其与AV接受的真正行为相互依赖性及其独特行为根源的了解。本研究通过引入递归的三变量计量经济模型来解决这些差距，该模型联合估计自动驾驶汽车的接受程度、感知到的安全问题和当前的出行行为（以年车辆行驶里程或VMT为代表）。递归结构解释了结构内生性，使安全关注和出行行为如何影响自动驾驶接受的无偏估计成为可能，同时将两者视为由其自身决定因素形成的内生结构。为了进一步增强行为现实性，该模型采用结构方程模型将潜在心理构念纳入其中。来自加州声明偏好数据集的实证结果强调，安全问题和潜在的车辆成本意识是自动驾驶汽车接受度的两个主要障碍，这表明建立信任活动和财政激励等政策可以刺激自动驾驶汽车的接受度。尽管对安全问题的关注程度较低，但行驶里程高的人对自动驾驶汽车的接受程度也较低，这表明，人们可能会习惯性地放弃驾驶控制权，这也挑战了人们的猜测，即用户喜欢车内节能，以及自动驾驶汽车促进了城市扩张。共享出行热情和潜在的车辆性能偏好缓解了自动驾驶汽车的安全担忧。性别和种族差异仍然存在，女性表达了更多的安全担忧，而亚洲人则表现出更高的AV接受度。

{"title":"Autonomous vehicle adoption behavior and safety concern: A study of public perception","authors":"Fatemeh Nazari , Mohamadhossein Noruzoliaee , Abolfazl (Kouros) Mohammadian","doi":"10.1016/j.multra.2025.100252","DOIUrl":"10.1016/j.multra.2025.100252","url":null,"abstract":"<div><div>Realizing the economic and societal benefits of autonomous vehicles (AVs) hinges on widespread public acceptance. However, existing research offers limited insights into two key behavioral factors shaping AV acceptance, namely, perceived AV safety concern and travel behavior, the latter reflecting how heterogenous mobility patterns influence the AV acceptance. These factors are often treated as exogenous, limiting insight into their true behavioral interdependencies with AV acceptance and their distinct behavioral roots. This study addresses these gaps by introducing a recursive trivariate econometric model that jointly estimates AV acceptance, perceived safety concern, and current travel behavior (proxied by annual vehicle-miles traveled or VMT). The recursive structure accounts for structural endogeneity, enabling the unbiased estimation of how safety concern and travel behavior influence AV acceptance, while treating both as endogenous constructs shaped by their own determinants. To further enhance behavioral realism, the model incorporates latent psychological constructs using structural equation modeling. Empirical results from a California stated preference dataset highlight that safety concern and latent vehicle cost consciousness are the two dominant deterrents to AV acceptance, suggesting that policies such as trust-building campaigns and financial incentives can stimulate AV acceptance. Despite showing less safety concern, high-VMT individuals exhibit lower AV acceptance, suggesting potential habitual inertia in ceding driving control and challenging conjectures that users embrace in-vehicle saving and that AVs promote urban sprawl. Shared mobility enthusiasm and latent vehicle performance preference alleviate AV safety concern. Gender and racial gaps persist, with women expressing greater safety concerns and Asians exhibiting higher AV acceptance.</div></div>","PeriodicalId":100933,"journal":{"name":"Multimodal Transportation","volume":"5 1","pages":"Article 100252"},"PeriodicalIF":0.0,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145420458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0