Transportation Research Part C-Emerging Technologies最新文献_第5页

Human-driven vehicles’ cruising versus autonomous vehicles’ back-and-forth congestion: The effects on traveling, parking, and congestion 人类驾驶汽车巡航与自动驾驶汽车来回拥堵：对出行、停车和拥堵的影响

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-09-04 DOI: 10.1016/j.trc.2025.105320

Xiaojuan Yu , Vincent A.C. van den Berg

{"title":"Human-driven vehicles’ cruising versus autonomous vehicles’ back-and-forth congestion: The effects on traveling, parking, and congestion","authors":"Xiaojuan Yu , Vincent A.C. van den Berg","doi":"10.1016/j.trc.2025.105320","DOIUrl":"10.1016/j.trc.2025.105320","url":null,"abstract":"<div><div>This paper explores the congestion interaction between human-driven vehicles (HVs) cruising for parking and autonomous vehicles (AVs) driving back and forth: first going into the city and then traveling back to park. To capture the spatial distribution of parking, we develop a continuous spatial optimization model, with a discrete choice logit model governing the choice between the two modes. Various congestion externalities are considered in the proposed model, including HVs’ cruising and searching for parking, as well as AVs’ back-and-forth travel. We analyze the joint travel mode and parking location choice equilibrium in the absence of pricing and in the social optimum that maximizes welfare. The social optimum is derived using optimal control method under user-equilibrium constraints. The introduction of AVs reduces parking demand in the city and results in a smaller city with lower parking density. Without pricing, the introduction of AVs may increase or decrease congestion, depending on if the effects of HVs cruising or AVs’ back-and-forth travel dominate. Thus, AVs may be underused or overused. With optimal pricing, the introduction of AVs always improves welfare. A stronger effect of AVs’ back-and-forth travel on congestion tends to increase the number of AVs and reduce the size of the city. This surprising result is because the effect of back-and-forth congestion may be larger for HVs than for AVs.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105320"},"PeriodicalIF":7.6,"publicationDate":"2025-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144989359","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A two-stage reservation and allocation approach for smart parking system considering parking unpunctuality 考虑停车不守时的智能停车系统两阶段预约与分配方法

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-09-01 DOI: 10.1016/j.trc.2025.105304

Xiaoyun Wang , Meng Xu , Haohan Xiao , Qiaoli Yang

{"title":"A two-stage reservation and allocation approach for smart parking system considering parking unpunctuality","authors":"Xiaoyun Wang , Meng Xu , Haohan Xiao , Qiaoli Yang","doi":"10.1016/j.trc.2025.105304","DOIUrl":"10.1016/j.trc.2025.105304","url":null,"abstract":"<div><div>To deal with the challenges caused by parking unpunctuality in current smart parking service management, this paper proposes a two-stage approach for the reservation and allocation problem. During the reservation stage, the expected service failure rates for accepted requests are calculated and a 0–1 quadratic programming (<strong>QP</strong>) model is proposed to generate a pre-matching scheme. During the allocation stage, a rolling horizon allocation approach is developed to allocate or reallocate parking spaces to accepted requests. For each allocation period, Dynamic Adjustment Strategies I-VI are employed to implement the pre-matching scheme with the real-time space occupancy and drivers’ location information, followed by the reallocation for failed requests. Furthermore, we obtain the representative distributions of parking arrival and duration for weekdays and weekends using the PH distribution, which can effectively capture their burstiness and clustering properties. The distributions of drivers’ arrival and departure time deviations are learned based on parking reservation data collected in Beijing. Results demonstrate that the proposed two-stage approach outperforms three baseline approaches (i.e., BDA, IRSA, and IASA) by reducing the service failure rate and increasing both theparking utilization rate and the platform’s net profit. The distributions of parking arrival and duration are shown to influence the performance evaluation of the proposed two-stage approach. Additionally, the improvements in the effectiveness of operational decisions during the allocation stage are more pronounced than those observed for tactical decisions during the reservation stage, particularly as the allocation frequency increases. Managerial implications for alleviating parking unpunctuality are further elaborated.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105304"},"PeriodicalIF":7.6,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144922505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

VLM-RL: A unified vision language models and reinforcement learning framework for safe autonomous driving VLM-RL：用于安全自动驾驶的统一视觉语言模型和强化学习框架

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-31 DOI: 10.1016/j.trc.2025.105321

Zilin Huang , Zihao Sheng , Yansong Qu , Junwei You , Sikai Chen

{"title":"VLM-RL: A unified vision language models and reinforcement learning framework for safe autonomous driving","authors":"Zilin Huang , Zihao Sheng , Yansong Qu , Junwei You , Sikai Chen","doi":"10.1016/j.trc.2025.105321","DOIUrl":"10.1016/j.trc.2025.105321","url":null,"abstract":"<div><div>In recent years, reinforcement learning (RL)-based methods for learning driving policies have gained increasing attention in the autonomous driving community and have achieved remarkable progress in various driving scenarios. However, traditional RL approaches rely on manually engineered rewards, which require extensive human effort and often lack generalizability. To address these limitations, we propose <strong>VLM-RL</strong>, a unified framework that integrates pre-trained Vision-Language Models (VLMs) with RL to generate reward signals using image observation and natural language goals. The core of VLM-RL is the contrasting language goal (CLG)-as-reward paradigm, which uses positive and negative language goals to generate semantic rewards. We further introduce a hierarchical reward synthesis approach that combines CLG-based semantic rewards with vehicle state information, improving reward stability and offering a more comprehensive reward signal. Additionally, a batch-processing technique is employed to optimize computational efficiency during training. Extensive experiments in the CARLA simulator demonstrate that VLM-RL outperforms state-of-the-art baselines, achieving a 10.5% reduction in collision rate, a 104.6% increase in route completion rate, and robust generalization to unseen driving scenarios. Furthermore, VLM-RL can seamlessly integrate almost any standard RL algorithms, potentially revolutionizing the existing RL paradigm that relies on manual reward engineering and enabling continuous performance improvements. The demo video and code can be accessed at: <span><span>https://zilin-huang.github.io/VLM-RL-website/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105321"},"PeriodicalIF":7.6,"publicationDate":"2025-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144919835","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Solving the Integrated-tasks satellite range scheduling problem: A surrogate-assisted variable neighborhood search approach 求解多任务卫星距离调度问题：一种代理辅助变量邻域搜索方法

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-30 DOI: 10.1016/j.trc.2025.105312

Yu Wang , Jingnan Dong , Xiaoming Xu , Hua Wang , Jiancheng Long

{"title":"Solving the Integrated-tasks satellite range scheduling problem: A surrogate-assisted variable neighborhood search approach","authors":"Yu Wang , Jingnan Dong , Xiaoming Xu , Hua Wang , Jiancheng Long","doi":"10.1016/j.trc.2025.105312","DOIUrl":"10.1016/j.trc.2025.105312","url":null,"abstract":"<div><div>This study addresses the large-scale Satellite Range Scheduling Problem (SRSP) that integrates Tracking, Telemetry, and Command (TT&C) and Data Transmission (DT) tasks (ISRSP), while considering the capabilities of ground stations in supporting satellite operations. By combining scheduling segments into available satellite operating arcs, we first formulate a binary integer programming model to maximize the fulfillment of two kinds of tasks (TT&C and DT). A comprehensive surrogate-assisted variable neighborhood search (SVNS) algorithm is then proposed to solve the ISRSP. We prove that the SVNS algorithm converges to the set of local optimal solutions with probability 1. Large-scale numerical experiments where the numbers of schedule tasks range from 5,000 to 10,000 are conducted to evaluate the effectiveness of the developed model and gauge the efficiency of the proposed algorithm. The results are compared with several state-of-the-art algorithms and demonstrate that the SVNS algorithm outperforms others in terms of convergence time, accuracy and solution stability.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105312"},"PeriodicalIF":7.6,"publicationDate":"2025-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144917329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Rebalancing an e-scooter-sharing system with en-route charging capability 重新平衡具有途中充电能力的电动滑板车共享系统

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-29 DOI: 10.1016/j.trc.2025.105295

Xiangyu Jin , Dong Han , Yufeng Cao , Yu Yang

{"title":"Rebalancing an e-scooter-sharing system with en-route charging capability","authors":"Xiangyu Jin , Dong Han , Yufeng Cao , Yu Yang","doi":"10.1016/j.trc.2025.105295","DOIUrl":"10.1016/j.trc.2025.105295","url":null,"abstract":"<div><div>The proliferation of e-scooter fleets complements urban transportation by providing flexibility and convenience. However, these fleets also face operational challenges, particularly in e-scooter rebalancing and battery recharging, which are crucial for maintaining desired service levels. This paper explores an innovative operational concept of en-route charging, where e-scooters are recharged while being transported between locations on rebalancing vehicles. When managed effectively, operations of rebalancing and recharging offer substantial benefits to e-scooter fleet operators and urban transportation users. We present a novel integrated decision model (IDM) that jointly addresses routing, waiting, and handling decisions to meet service-level requirements and maximize profit. To tackle computational challenges, we decompose decisions into two stages and prove the total unimodularity (TU) of the second-stage problem’s coefficient matrix. This property allows us to simplify the second-stage integer program to a manageable linear program without sacrificing solution quality. An extensive numerical study based on a real-world shared e-scooter system in Minneapolis, leveraging the TU property, demonstrates computational efficiency improvements of 2.68 times and 4.95 times with 20 and 30 gathering points, respectively. Meanwhile, the proposed model outperforms the two alternative baseline algorithms, increasing operational profits by 70%. Our study also quantifies the economic benefits of en-route charging, showing an average total profit increase of $238.89 (an 8.9% increase) compared to a centralized recharging approach. Furthermore, we address the equity aspect of operations, showing that enforcing fairness requirements results in a negligible average profit loss of $2.03 (0.07%) of the daily operations. These findings provide valuable insights for both system operators and policymakers, highlighting the operational and economic advantages of en-route charging in e-scooter fleet management.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105295"},"PeriodicalIF":7.6,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A two-round broadcasting matching mechanism in ride-sourcing markets: Implication and optimization 网约车市场的两轮广播匹配机制：启示与优化

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-28 DOI: 10.1016/j.trc.2025.105318

Xiaoran Qin , Hai Yang , Yuhan Liu

{"title":"A two-round broadcasting matching mechanism in ride-sourcing markets: Implication and optimization","authors":"Xiaoran Qin , Hai Yang , Yuhan Liu","doi":"10.1016/j.trc.2025.105318","DOIUrl":"10.1016/j.trc.2025.105318","url":null,"abstract":"<div><div>The rapid growth of the ride-sourcing market has intensified competition, with driver–passenger matching strategies playing a pivotal role in platform performance. To attract more supply resources, some platforms adopt an order-broadcasting matching mechanism with a fixed broadcasting radius. Although this mechanism provides drivers more flexibility, it often struggles to balance matching probability and passenger wait times, leading to inefficiencies. In this regard, we propose an adaptive two-round broadcasting matching mechanism that dynamically adjusts the search radius based on drivers’ responses. This mechanism attempts to improve both the driver’s flexibility and the passenger experience but introduces new challenges in determining optimal configurations over multiple matching rounds. To tackle these challenges, we characterize the matching process and general results of the two-round broadcasting mechanism, particularly incorporating the heterogeneous acceptance behaviors of drivers and passengers under this special scheme. These models enable us to explore the impact of decision variables (i.e., radius in each round and time intervals) on system performance, such as matching probability and expected waiting and pickup time. Moreover, we derive the conditions which our proposed mechanism outperforms the traditional fixed-radius one. Our numerical results show that the platform can achieve better matching outcomes by appropriately adjusting its two-round matching radii and time intervals under different demand–supply conditions.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105318"},"PeriodicalIF":7.6,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Equitable port state control in maritime transportation: A data-driven optimization approach 海上运输中港口国的公平控制：数据驱动的优化方法

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-28 DOI: 10.1016/j.trc.2025.105303

Yanxia Guan, Xuecheng Tian, Yiwei Wu, Shuaian Wang

{"title":"Equitable port state control in maritime transportation: A data-driven optimization approach","authors":"Yanxia Guan, Xuecheng Tian, Yiwei Wu, Shuaian Wang","doi":"10.1016/j.trc.2025.105303","DOIUrl":"10.1016/j.trc.2025.105303","url":null,"abstract":"<div><div>Port state control (PSC) inspection, an important measure to prevent maritime accidents, prioritizes inspections for high-risk ships. The deficiency value of a ship is commonly used to quantify its potential risk, and the predicted deficiency value, derived from a prediction model trained through historical inspection records, is used to assist ship selection. Single prediction models are commonly used to estimate ships’ deficiency values. However, their performance can be affected by noise interference, which can decrease the stability and robustness of models. Ensemble learning is an effective method to enhance the prediction performance by combining results from multiple models. However, despite their advantages, current ensemble learning methods face challenges in effectively distinguishing ships with similar deficiency values, which can impact the effectiveness and fairness of the decision making. To address this issue, we propose a novel heterogeneous ensemble learning method that combines distributional estimation with lexicographic optimization to enhance ship selection performance. The model first constructs a comparison matrix between ships by distributional estimation. Subsequently, a lexicographic optimization model is applied to construct a discriminative weight vector based on the comparison matrix that assigns larger weights to higher-risk ships. Finally, we conduct a comprehensive evaluation of the effectiveness and fairness of the proposed method. We consider five base prediction models and select three widely used ensemble learning methods—averaging, stacking, and boosting—for further comparison. Experimental results from multiple test scenarios on real-world data from the port of Hong Kong demonstrate that the proposed method outperforms other methods in terms of both effectiveness and fairness.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105303"},"PeriodicalIF":7.6,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Decision-focused learning for optimal subsidy allocation in ride-hailing services 基于决策的网约车补贴最优分配学习

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-28 DOI: 10.1016/j.trc.2025.105301

Jiaqi Yang , Lexiao Chen , Zicheng Su , Wanjing Ma , Zhichao Zou , Kun An

{"title":"Decision-focused learning for optimal subsidy allocation in ride-hailing services","authors":"Jiaqi Yang , Lexiao Chen , Zicheng Su , Wanjing Ma , Zhichao Zou , Kun An","doi":"10.1016/j.trc.2025.105301","DOIUrl":"10.1016/j.trc.2025.105301","url":null,"abstract":"<div><div>Effective subsidy strategies are essential for ride-hailing platforms. The key is to accurately predict trip completion rates before determining the optimal subsidy decision. However, the classic predict-then-optimize framework often falls short due to discrepancies between the prediction model, which minimizes fitting error, and the decision model, which minimizes decision error. To tackle the above issues, this paper presents a decision-focused learning (DFL) approach for optimal subsidy allocation in ride-hailing services at the city level. First, we formulate the subsidy allocation as a bi-level optimization problem, in which the lower-level individual distribution can be approximated via a threshold-based strategy. Moreover, the effect of the subsidy is explicitly modeled, which considers the combined impacts of both pricing and matching aspects. Second, the prediction model for the completion rate is trained to directly minimize downstream decision loss, integrating the prediction module and optimization task in an end-to-end manner. To address the challenge of conducting gradient backpropagation from the decision model to the prediction model, we utilize a surrogate decision-focused loss function, whose convexity in the context of the subsidy allocation problem is theoretically proven. Third, we develop a decision-focused fine-tuning mechanism to handle the non-differentiable prediction model, allowing any upstream prediction model to be adjusted based on the downstream decision loss. The proposed DFL framework is tested on real-world data from DiDi Chuxing. Results show that decision-focused learning can increase platform revenue by 1.51% compared to traditional predict-then-optimize solutions, as the prediction model effectively learns the weights of different cities in the decision objective.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105301"},"PeriodicalIF":7.6,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913786","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic truck-drone collaborative transportation during hurricanes 飓风期间的动态卡车-无人机协同运输

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-28 DOI: 10.1016/j.trc.2025.105322

Keyu Li , Haipeng Cui , Qiang Meng , Xuxin Zhang

{"title":"Dynamic truck-drone collaborative transportation during hurricanes","authors":"Keyu Li , Haipeng Cui , Qiang Meng , Xuxin Zhang","doi":"10.1016/j.trc.2025.105322","DOIUrl":"10.1016/j.trc.2025.105322","url":null,"abstract":"<div><div>The increasing frequency and severity of extreme weather events highlight the urgent need for resilient and adaptive logistics solutions. Existing models often overlook the complexities introduced by extreme weather conditions, which can significantly impair the operational efficiency of transportation means including trucks and drones. To address this gap, we propose a dynamic collaborative truck-drone vehicle routing problem (VRP) with mixed line hauls and back hauls in wind-varying conditions (CVRPLB-WV) under hurricanes lasting three to five days, aiming to maintain transportation resilience by reducing delivery delays and maximizing profit. The proposed CVRPLB-WV leverages a flexible truck-drone collaboration strategy to adapt to varying wind speeds by restricting available transportation resources and speed limits. We formulate the CVRPLB-WV problem as a mixed integer programming (MIP) model and solve the problem by proposing a resilient adaptive large neighborhood search (R-ALNS) algorithm under the rolling horizon (RH) framework to handle the inherent complexities. Extensive numerical experiments utilizing data from three real Hurricane events, i.e., Nalgae, Koinu, and Sura, demonstrate that our approach consistently outperforms the other benchmark algorithms (ALNS, LNS, Greedy, and reinforcement learning algorithms) by achieving up to 79.08 % delay reduction and higher profits. Additionally, a sensitivity analysis explores the impact of varying truck-to-drone ratios, drone endurance, capacities, hurricane durations, and collaboration strategies, providing valuable managerial insights that underscore the practical applicability and robustness of our approach.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105322"},"PeriodicalIF":7.6,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Learning universal human mobility patterns with a foundation model for cross-domain data fusion 使用跨领域数据融合的基础模型学习通用的人类迁移模式

IF 7.6 1区工程技术

Transportation Research Part C-Emerging Technologies Pub Date : 2025-08-28 DOI: 10.1016/j.trc.2025.105311

Haoxuan Ma , Xishun Liao , Yifan Liu , Qinhua Jiang , Chris Stanford , Shanqing Cao , Jiaqi Ma

{"title":"Learning universal human mobility patterns with a foundation model for cross-domain data fusion","authors":"Haoxuan Ma , Xishun Liao , Yifan Liu , Qinhua Jiang , Chris Stanford , Shanqing Cao , Jiaqi Ma","doi":"10.1016/j.trc.2025.105311","DOIUrl":"10.1016/j.trc.2025.105311","url":null,"abstract":"<div><div>Human mobility modeling is critical for urban planning and transportation management, yet existing approaches often lack the integration capabilities needed to handle diverse data sources. We present a foundation model framework for universal human mobility patterns that leverages cross-domain data fusion and large language models to address these limitations. Our approach integrates multi-modal data of distinct nature and spatio-temporal resolution, including geographical, mobility, socio-demographic, and traffic information, to construct a privacy-preserving and semantically enriched human travel trajectory dataset. Our framework demonstrates adaptability through domain transfer techniques that ensure transferability across diverse urban contexts, as evidenced in case studies of Los Angeles (LA) and Egypt. The framework employs LLMs for semantic enrichment of trajectory data, enabling comprehensive understanding of mobility patterns. Quantitative evaluation shows that our generated synthetic dataset accurately reproduces mobility patterns observed in empirical data. The practical utility of this foundation model approach is demonstrated through large-scale traffic simulations for LA County, where results align well with observed traffic data. On California’s I-405 corridor, the simulation yields a Mean Absolute Percentage Error of 5.85% for traffic volume and 4.36% for speed compared to Caltrans PeMS observations, illustrating the framework’s potential for intelligent transportation systems and urban mobility applications.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"180 ","pages":"Article 105311"},"PeriodicalIF":7.6,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144913648","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0