Transportation Research Part C-Emerging Technologies最新文献

{"title":"Toward low-burden travel survey: Identifying travel modes from GPS tracks fusing individual histories and enumerated annotations","authors":"Jiaqi Zeng ,&nbsp;Yulang Huang ,&nbsp;Meng Zhang ,&nbsp;Wenbin Yao ,&nbsp;Sheng Jin ,&nbsp;Dianhai Wang","doi":"10.1016/j.trc.2024.104975","DOIUrl":"10.1016/j.trc.2024.104975","url":null,"abstract":"<div><div>GPS-based travel surveys, coupled with automatic Travel Mode Identification (TMI) techniques, have emerged as effective tools for capturing travel details while reducing participant annotation burdens. However, existing models typically identify travel modes on a per-trip basis without considering individual travel regularity. More importantly, incorrect identification increases modification burdens, an issue that has not been adequately assessed or addressed. We propose a novel TMI scheme for travel surveys: participants enumerate their used travel modes before TMI, and the History and Annotation Informed Network (HAIN) model integrates individual historical travel information and enumerated annotations to infer the current travel mode chain, thereby improving accuracy and reducing the need for modifications. The designed history and annotation fusion modules in HAIN are plug-and-play and can operate separately. Additionally, we introduce the “number of edits” to quantitatively assess user annotation burden. We divide data by travelers and conduct three-fold cross-validation to approximate real-world scenarios. Results show that the accuracy of the state-of-the-art model is 83.14%; it reaches 86.3% when identifying trips sequentially with previously inferred histories; using corrected histories improves accuracy to 88.26%; and incorporating enumerated annotations raises it to 96.03%. Correspondingly, the annotation burdens are reduced to 43.8%, 39.9%, 36.6%, and 24.3% of what they would be without TMI. The two fusion modules also enhance the performance of baseline models. The history fusion improves model robustness when incorrect annotations occur. Comprehensive experiments indicate that the proposed scheme significantly enhances data collection efficiency and improves user experience.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104975"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096498","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}

{"title":"Enhancing public transit adoption through personalized incentives: a large-scale analysis leveraging adaptive stacking extreme gradient boosting in China","authors":"Haoyang Yan ,&nbsp;Xiaolei Ma ,&nbsp;Bing Liu ,&nbsp;Erlong Tan ,&nbsp;Yujie Li ,&nbsp;Zirui Ni ,&nbsp;Tian-Liang Liu","doi":"10.1016/j.trc.2024.104992","DOIUrl":"10.1016/j.trc.2024.104992","url":null,"abstract":"<div><div>Motivating individuals to utilize public transportation through financial strategies, including both rewards and penalties, has been acknowledged as an effective approach to manage traffic demand and mitigate congestion-related issues. Personalized travel rewards, in contrast to economic sanctions like road tolls, tend to be more socially accepted. Nonetheless, insights into the effectiveness of personalized incentives remain limited, often constrained by studies that rely on small, non-representative samples of travelers. This study seeks to identify the variables that prompt individuals to switch to public transportation, drawing on extensive quasi-experimental data from a widespread public transit incentive program featured in one of China’s the largest navigation apps. This data encompasses the sociodemographic details of users, as well as their local and long-distance travel patterns. Both a binary Logit model and an adaptive stacking extreme gradient boosting (AS-XGB) model are applied to interpret and predict the changes in users’ public transit usage. Besides gender, job type and preferred travel mode, incentive reward category is found to be one of the significant determinants. In particular, rewards such as breakfast bread or travel vouchers have proven more effective than other types of incentives, like supermarket coupons or tissue gift bags. Female participants, individuals without children, and those who used public transportation in the week prior to receiving the incentives showed a higher propensity to embrace these rewards. However, the influence of education level, car ownership status, or preferred travel mode largely varies as the city’s development level. For intercity travel, regardless of whether the user owns a car or not, her/his income level and education level both have significant impacts on the incentive effectiveness.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104992"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143104805","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}

{"title":"Cost-competitiveness analysis of mobile chargers in an electric vehicle parking and charging system","authors":"Yanling Deng ,&nbsp;Zhibin Chen ,&nbsp;Xi Lin","doi":"10.1016/j.trc.2024.104951","DOIUrl":"10.1016/j.trc.2024.104951","url":null,"abstract":"<div><div>Currently, mobile chargers (MCs) are gaining popularity owing to their flexibility and the potential to solve the widespread issue of chargers being occupied after the charging process has finished, due to delayed departures of electric vehicles (EVs). In this study, we investigate the adoption of MCs in the EV parking and charging system (EVPCS) and demonstrate its cost-competitiveness through comparison with fixed chargers (FCs). First, we propose a modified M/M/n/K queueing model with two-phase services to capture the EVs’ dwell-after-charging behavior. Then we present the steady-state analysis with a matrix analytical method to analyze the properties of the proposed models. To evaluate and compare the performances of these two types of charging facilities, several key measures like blocking probability, average queue length and delay, and chargers’ utilization rate are derived, and extensive experiments exploring diverse scenarios are obtained. Numerical results uncover that: (1) the EVPCS configuration and EVs’ arrival and service rates have distinct impacts on the performance metrics of different types of chargers; (2) the MC may lose its competitiveness if the EV arrival rate is relatively low along with a high charging service rate or when the charger proportion approaches 1 in a small-sized EVPCS; only when the system is overloaded would the MC have a higher level of service than the FC; (3) in terms of cost efficiency, MCs demonstrate better competitiveness than the low-powered FCs considering the equivalent service with higher consumer surplus, while competing with moderately to high-powered FCs, MCs have little superiority; MCs are more profitable to the operator when competing with either low or high-powered FCs, but not for moderately-powered FCs; yet, the MC’s charging powers is required to be at an acceptably moderate level to demonstrate its cost-competitiveness. Furthermore, analytical formulas have been developed to approximate the two-phase queueing model under certain scenarios, and their accuracy has been compared with the customized two-phase queueing model.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104951"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143095601","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}

{"title":"Multi-agent trajectory prediction at unsignalized intersections: An improved generative adversarial network accounting for collision avoidance behaviors","authors":"Lei Zhao,&nbsp;Wei Zhou,&nbsp;Sixuan Xu,&nbsp;Yuzhi Chen,&nbsp;Chen Wang","doi":"10.1016/j.trc.2024.104974","DOIUrl":"10.1016/j.trc.2024.104974","url":null,"abstract":"<div><div>Accurate trajectory prediction for multiple agents (i.e., vehicles, bicyclists, and pedestrians) is the premise of launching proactive interventions, which can serve as an effective way to improve traffic safety at unsignalized intersections. The distinctive characteristic of unsignalized intersections lies in their disorderly traffic organization, prompting traffic agents to be extra vigilant towards other agents to prevent collisions. As such, the primary focus of multi-agent trajectory prediction lies in acquiring a deep understanding of their interactive behavior patterns when encountering potential collisions. To achieve this, this study proposes an improved generative adversarial network (GAN) that can properly model collision avoidance behaviors of multiple agents when predicting their trajectories. Specifically, attention pooling modules are employed to capture interactions among multiple agents. A graph convolution network (GCN) based collision extraction module is applied to identify potential collisions and model the collision avoidance behaviors of traffic agents. Experimental results on <em>InD</em> dataset demonstrate that the proposed framework attained a more accurate and reliable performance for multi-agent trajectory prediction. In different interactive scenarios, such as when vehicles yield or don’t yield, the results illustrated via the Distance-velocity (DV) diagram display a significant level of robustness. Furthermore, the conflict points, count of dangerous interactions, and Post-Encroachment Time, as computed from these predicted trajectories, also align well with the ground truth. This indicates that the proposed framework effectively captures the pattern of collision avoidance behaviors of multiple agents, which has potential to serve as an effective way to enhance traffic safety at unsignalized intersections.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104974"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096022","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}

{"title":"Physics informed deep generative model for vehicle trajectory reconstruction at arterial intersections in connected vehicle environment","authors":"Mengyun Xu ,&nbsp;Jie Fang ,&nbsp;Prateek Bansal ,&nbsp;Eui-Jin Kim ,&nbsp;Tony Z. Qiu","doi":"10.1016/j.trc.2024.104985","DOIUrl":"10.1016/j.trc.2024.104985","url":null,"abstract":"<div><div>Inferring the complete traffic flow time–space diagram using vehicle trajectories provides a holistic perspective of traffic dynamics at intersections to traffic managers. However, obtaining all vehicle trajectories on the road is infeasible. To this end, a novel framework that combines the conditional deep generative model and physics-based car-following model is proposed to reconstruct all vehicle trajectories from sparsely available connected vehicle (CV) trajectories at the intersection. The proposed framework has two novel components: Arrival Generative Adversarial Network (Arrival-GAN) and Trajectory-GAN. The Arrival-GAN reproduces stochastic vehicle arrival patterns by considering the interaction between adjacent intersections (e.g., signal control scheme) and the interaction between multiple vehicles from historical vehicle trajectories, circumventing the conventionally adopted unrealistic assumptions of uniform vehicle arrivals. The Trajectory-GAN model takes the baseline trajectory deduced by the physics-based car-following model as prior information and refines it by dynamically adapting driving behavior in response to the varying traffic conditions in a data-driven manner. This hybrid approach leverages the advantages of data-driven (i.e., flexibility) and theory-driven approaches (i.e., interpretability) complementarily. The proposed framework outperforms conventional benchmark models in the simulated arterial network and the real-world datasets, reconstructing a complete time–space diagram at intersections with markedly enhanced accuracy, particularly in low-traffic-density scenarios. This study showcases the potential of utilizing CV data and physics-informed deep learning to improve our understanding of traffic dynamics, empowering traffic managers with novel insights for efficient intersection management.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104985"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096047","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}

{"title":"Routing of autonomous vehicles for optimal total system cost under equilibrium","authors":"Faizan Ahmad Kashmiri,&nbsp;Hong K. Lo,&nbsp;Yue Huai,&nbsp;Zheng Liang,&nbsp;Lubing Li","doi":"10.1016/j.trc.2025.105004","DOIUrl":"10.1016/j.trc.2025.105004","url":null,"abstract":"<div><div>The notion of optimal mobility via Autonomous Vehicles (AVs) controlled by a transportation management centre (TMC) has received considerable attention in the past few years. One of the novel demand management strategies entails Average Travel Time (AVT) equilibrium, a cyclic path assignment that travellers experience at the end of a fixed cycle (a period of many days) while maintaining system optimal (SO) path flows every day. However, this work is restricted to smaller networks because of the intricate solution strategies involved. To address large networks in notably lower run times, we thus put forth a linear-based AVT approach. Furthermore, applying such TMCs is limited to system time only, with system emissions receiving negligible attention. In this regard, we propose a TMC that preserves an optimal total system cost (OTSC), including travel time and emission. Unlike earlier AVT works, the average vehicle travel time and emission cost (AVTEC) equilibrium is converted from a non-linear complementarity problem to a linear problem by manipulating the inherent structure of the problem. Similar to AVT, when a fixed cycle ends, travellers from the same Origin-Destination (OD) pair will have taken distinct routes within a given path combination set in accordance with predetermined time proportions and experience equal and minimal AVTEC. Numerical investigations demonstrate that, in contrast to previous approaches to AVT, the novel linear AVTEC framework can solve extended networks significantly faster. We also investigate the mixed equilibrium (ME) scenario, in which specific travellers (referred to as non-subscribers) who choose not to subscribe to the TMC but adhere to their user equilibrium (UE) routing patterns and compete with the OTSC travellers. Results indicate that the non-subscribers will outperform the subscribers, giving non-subscribers incentives to not join the TMC. To deter this from occurring, we determine tolls upon non-subscribers/UE travellers based on the travel cost difference (TCD) between the AVTEC of TMC subscribers and the minimum UE travel cost of non-subscribers. While imposing tolls upon UE travellers, we consider their heterogeneity through their value-of-time (VOT) distribution to equalize the minimum private costs of non-subscribers with the AVTEC of subscribers. In addition, we propose link-based and sensitivity-based equilibrium strategies regarding ME, tolls and different penetration rates of subscribers and non-subscribers based on their VOT distribution to solve extensive networks.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 105004"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096392","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}

{"title":"Single-stage train formation in railway marshaling stations under an extended railcar-to-track assignment policy","authors":"Siyu Zhang ,&nbsp;Jun Zhao ,&nbsp;Bojian Zhang ,&nbsp;Jiaxi Zhao ,&nbsp;Qiyuan Peng","doi":"10.1016/j.trc.2024.104972","DOIUrl":"10.1016/j.trc.2024.104972","url":null,"abstract":"<div><div>This paper studies a single-stage train formation problem in railway marshaling stations, aiming to efficiently assign railcars to classification tracks with one roll-in and one pull-out operation for ensuring the formation of outbound trains. Assigning railcars to classification tracks by blocks (block-to-track), by outbound trains (train-to-track), and by need (railcar-to-track) are three typical policies widely addressed in the literature. An extended railcar-to-track policy is investigated by combining the first and third policies, where railcars are preferentially assigned to their fixed-use classification tracks through a specified block-to-track scheme and then to other classification tracks if necessary, while re-humping and re-sorting operations are eliminated. We formulate the formation problem under this policy as a binary linear programming model with the objective of minimizing the total weighted cost required for train formation, including both the weighted roll-in cost and the weighted pull-out cost. A two-phase decomposition algorithm, which divides our model into a roll-in and a pull-out subproblem, is developed to improve the solving efficiency. For the roll-in subproblem, a novel group-indexed model is constructed to determine a railcar-to-track scheme with minimal total weighted roll-in cost and simplified pull-out cost. For the following pull-out subproblem, the objective is to determine a pull-out scheme that minimizes the total weighted pull-out cost. This subproblem is decomposed further into multiple simplified problems, each of which is formulated into a quadratic assignment model for each outbound train, enabling rapid solving times of this subproblem. Computational results on a set of realistic instances reveal that our algorithm outperforms two benchmark approaches, in which the roll-in subproblem is formulated respectively as a big-M and an arc-indexed model inspired by existing models, and an imitated empirical approach used in practice. The potential superiority of our proposed policy to the three existing policies is also numerically validated.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104972"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096504","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}

{"title":"Prediction of traffic state variability with an integrated model-based and data-driven Bayesian framework","authors":"Xinyue Wu ,&nbsp;Andy H.F. Chow ,&nbsp;Wei Ma ,&nbsp;William H.K. Lam ,&nbsp;S.C. Wong","doi":"10.1016/j.trc.2024.104953","DOIUrl":"10.1016/j.trc.2024.104953","url":null,"abstract":"<div><div>Deriving statistical description of uncertainties associated with prediction of traffic states is essential to development of reliability-based intelligent transportation systems. This paper presents a Bayesian learning approach framework for predicting evolution of both traffic states and the associated variability. The proposed framework ensures the interpretability and stability of the predictions with an underlying state space model, and captures sophisticated dynamics of traffic variability via a data-driven recurrent neural network component. By maintaining the filtering structure in the specialized neural network component, the proposed integrated model overcomes the key limitations of deep learning systems by improving the data efficiency and providing interpretability. The framework is trained with a multivariate Gaussian negative log-likelihood loss function for quantifying both model and stochastic uncertainties. It is implemented and tested with actual traffic data collected from a Hong Kong Strategic Route. The case study shows that the proposed prediction framework can simultaneously retain the interpretability of the results while capture the complex dynamics of the evolution of traffic variability with the recurrent neural network component. This study contributes to the development of reliability-based intelligent transportation systems through the use of advanced statistical modeling and deep learning methods.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104953"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143095591","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}

{"title":"Multi-node joint optimization for fine-grained vehicle trajectory reconstruction using vehicle appearance and identity data","authors":"Mingkai Qiu ,&nbsp;Yuhuan Lu ,&nbsp;Xiying Li","doi":"10.1016/j.trc.2024.104995","DOIUrl":"10.1016/j.trc.2024.104995","url":null,"abstract":"<div><div>Automatic vehicle identification (AVI) data provides information about vehicles’ identity and location, which enables accurate vehicle trajectory extraction. However, collecting complete and continuous long-term trajectory data is challenging due to limitations in coverage rate and identification accuracy. To recover the incomplete trajectory, current research mainly focuses on coarse-grained trajectory reconstruction. While these methods are adept at reconstructing the road segments traversed by vehicles, they cannot restore the spatial–temporal details of vehicle journeys and capture individual variability. To address these limitations, we propose a Multi-node Joint Optimization (MNJO) model, which utilizes AVI data and individual appearance features to achieve fine-grained vehicle trajectory reconstruction. The MNJO comprises two stages: local instance association and global trajectory reconstruction. In the local instance association stage, we design an inter-vehicle bidirectional optimization mechanism, which integrates the competitive and associative interactions among vehicles to improve their association across different nodes. In the global trajectory reconstruction stage, we propose a trajectory optimization network for trajectory scoring based on the spatial–temporal characteristics and the local association results of all nodes along the trajectory. Due to the lack of related datasets, we construct the Regional Vehicle Information (RVI) dataset, the first for fine-grained trajectory reconstruction, collected from real-world AVI systems. Extensive experiments on the RVI show that the MNJO can achieve significant enhancement in reconstruction accuracy compared to other methods, demonstrating the effectiveness and superiority of the proposed method. To the best of our knowledge, this is the first study on fine-grained vehicle trajectory reconstruction.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104995"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096024","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}

{"title":"Hierarchical prediction uncertainty-aware motion planning for autonomous driving in lane-changing scenarios","authors":"Ruoyu Yao ,&nbsp;Xiaotong Sun","doi":"10.1016/j.trc.2024.104962","DOIUrl":"10.1016/j.trc.2024.104962","url":null,"abstract":"<div><div>Autonomous vehicles (AVs) are expected to achieve safe and efficient interactions with surrounding dynamic objects. Multi-lane driving scenarios, however, intensify the complexity of AV navigation, given the uncertainties associated with neighboring vehicles’ lane-changing intentions and the subsequent travel trajectories. Deep learning has demonstrated effectiveness in unraveling complex motion patterns, enabling stochastic predictions of intentions and trajectories. Nonetheless, reduced performance of deep-learning-based prediction may be observed in unseen driving environments owing to their “black-box” nature, potentially leading to incorrect decision-making in AV navigation in these environments. To address these challenges, this paper proposes a comprehensive AV planning framework that integrates hierarchical behavior prediction via deep learning with motion planning based on dynamic programming. A set of safety criteria is introduced within the motion planning module to accommodate hierarchical uncertainties in behavior patterns, adjustable based on the reliability of the prediction model and eschewing rigid distributional assumptions. An improved constrained iterative linear quadratic regulator is devised to handle the corresponding non-convex constraints and to offer efficient online solutions for AV navigation. Evaluations conducted with the INTERACTION and HighD datasets demonstrate the effectiveness of uncertainty-aware planning in enhancing AV safety performance.</div></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":"171 ","pages":"Article 104962"},"PeriodicalIF":7.6,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143096499","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}