IEEE Transactions on Intelligent Vehicles最新文献

The Transactions on Intelligent Vehicles Information 智能车辆信息学报

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-10-09 DOI: 10.1109/TIV.2026.3740250

引用次数: 0

Enhancing 3D LiDAR Object Detection in Real-World Infrastructure With Trained Models Using Simulation Data 利用仿真数据训练模型增强现实世界基础设施中的3D激光雷达目标检测

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-16 DOI: 10.1109/TIV.2026.3734950

Shunsuke Ito;Chaoran Zhao;Takuya Azumi

{"title":"Enhancing 3D LiDAR Object Detection in Real-World Infrastructure With Trained Models Using Simulation Data","authors":"Shunsuke Ito;Chaoran Zhao;Takuya Azumi","doi":"10.1109/TIV.2026.3734950","DOIUrl":"https://doi.org/10.1109/TIV.2026.3734950","url":null,"abstract":"The rapid development of autonomous driving technology has increased the need for accurate vehicle environmental perception, with infrastructure 3D LiDAR playing a crucial role in enhancing traffic safety and management. However, processing its large-scale, sparse, and noisy point clouds remains challenging, compounded by the high cost of annotation. This paper proposes a deep learning (DL)-based framework integrating foreground-background separation, noise modeling, and automatic point-cloud annotation. The DL-based separation method removes approximately 99% of background points, significantly improving object detection accuracy. The noise modeling approach transfers ray-wise geometric perturbations learned from real measurements to simulator-generated point clouds under a fixed LiDAR configuration, achieving up to 95% similarity based on normalized Chamfer Distance in the evaluated setting. This enables models trained solely on simulated data to reach 69% accuracy on real-world data at an IoU threshold of 0.5. Additionally, the proposed scenario framework integrates LiDAR point clouds with OpenSCENARIO 2.0 to support simulator-based automatic annotation and training-data generation. Experimental results demonstrate the effectiveness of the proposed framework for foreground reduction and sim-to-real training-data refinement in the evaluated infrastructure LiDAR setting. The present study does not claim validated robustness to adverse weather or broad vehicle-geometry variation, which remain important directions for future evaluation.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1210-1225"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11694244","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149131055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-Informed Digital Twins for Condition Monitoring of Autonomous Marine Vessels 自主船舶状态监测的物理信息数字孪生

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-17 DOI: 10.1109/TIV.2026.3733473

Agus Hasan;Sam Jackson

{"title":"Physics-Informed Digital Twins for Condition Monitoring of Autonomous Marine Vessels","authors":"Agus Hasan;Sam Jackson","doi":"10.1109/TIV.2026.3733473","DOIUrl":"https://doi.org/10.1109/TIV.2026.3733473","url":null,"abstract":"This paper presents the development of physics-informed digital twins for condition monitoring of autonomous marine vessels, with a focus on comprehensive coverage that includes the diagnosis of sensors, actuators, and internal systems. To enable accurate fault diagnosis, we introduce two physics-based algorithms, one stochastic and one deterministic, that estimate the magnitude of system faults. These algorithms leverage underlying physical principles to identify deviations from nominal behavior and provide interpretable diagnostic outputs. The proposed methods are benchmarked against simulation-based inference techniques enhanced with deep learning, which have gained popularity due to their ability to model complex patterns from large datasets. All proposed approaches are evaluated using a high-fidelity digital twin of a research platform, parameterized with experimentally identified hydrodynamic coefficients. The results demonstrate that the physics-informed algorithms consistently outperform the deep learning-based method in terms of diagnostic accuracy, particularly in scenarios with limited training data or where model interpretability is essential.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1085-1095"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149131074","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

Multi-Attribute Scene Identification for Autonomous Vehicles: Acquiring and Accumulating Knowledge From Diverse Datasets 自动驾驶汽车多属性场景识别：从不同数据集获取和积累知识

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-14 DOI: 10.1109/TIV.2026.3733981

Ke Li;Chenyu Zhang;Yuxin Ding;Xianbiao Hu;Ruwen Qin

{"title":"Multi-Attribute Scene Identification for Autonomous Vehicles: Acquiring and Accumulating Knowledge From Diverse Datasets","authors":"Ke Li;Chenyu Zhang;Yuxin Ding;Xianbiao Hu;Ruwen Qin","doi":"10.1109/TIV.2026.3733981","DOIUrl":"https://doi.org/10.1109/TIV.2026.3733981","url":null,"abstract":"Driving scenes are inherently heterogeneous and dynamic. Multi-attribute scene identification, as a high-level visual perception capability, provides autonomous vehicles (AVs) with essential contextual awareness to understand, reason through, and interact with complex driving environments. Although scene identification is best modeled as a multi-attribute classification problem through multitask learning, it faces two major challenges: the difficulty of acquiring balanced, comprehensively annotated datasets and the need to re-annotate all training data when new attributes emerge. To address these challenges, this paper introduces a novel deep learning method that integrates Knowledge Acquisition and Accumulation (KAA) with Consistency-based Active Learning (CAL). KAA leverages monotask learning on heterogeneous datasets with single-attribute annotations to build a knowledge foundation, while CAL bridges the marginal-to-joint distribution gap, adapting the foundation model to multi-attribute scene identification. An ablation study on the newly developed Driving Scene Identification (DSI) dataset demonstrates a 56.1% improvement over an ImageNet-pretrained baseline. Moreover, KAA-CAL outperforms state-of-the-art multi-attribute classification methods on the BDD100K and HSD datasets, using 85% fewer labeled training images during adaptation and successfully recognizing attributes not included in KAA training.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1143-1158"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149133838","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

IEEE Transactions on Intelligent Vehicles Publication Information IEEE智能车辆学报出版信息

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-10-09 DOI: 10.1109/TIV.2026.3740248

引用次数: 0

Target Point Selection in Path Following: A Review for Autonomous Vehicle Control 路径跟踪中的目标点选择：自动驾驶汽车控制研究综述

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-16 DOI: 10.1109/TIV.2026.3734841

Anis Koliai;Stephane Bazeille;Michel Basset;Rodolfo Orjuela

{"title":"Target Point Selection in Path Following: A Review for Autonomous Vehicle Control","authors":"Anis Koliai;Stephane Bazeille;Michel Basset;Rodolfo Orjuela","doi":"10.1109/TIV.2026.3734841","DOIUrl":"https://doi.org/10.1109/TIV.2026.3734841","url":null,"abstract":"Path following is a critical component of autonomous vehicle navigation, significantly contributing to the vehicle’s ability to maintain a desired path. This paper presents a comprehensive study on the design of path following control, which is structured in three parts: tracking errors, target point, controller design. The first part defines the references frames and the procedure for computing tracking errors. Next, a comprehensive review of the most commonly used methods for selecting target points is provided. These methods are studied in open loop by means of an analytical example. The open loop simulation results reveal that some methods can be used for control purposes, others for assessment purposes, some are unsuitable altogether. Building on this review, a proposed algorithm is used to construct four control models in order to design Linear Parameter-Varying (LPV) output feedback controllers. These controllers are simulated in high fidelity simulator. Closed-loop results, suggest that all methods have different convergence and tracking behaviors in low and high speeds. The target point can be considered as a design degree of freedom that can enhance or reduce the performance of path following controllers. Finally, this survey serves as a resource for researchers and practitioners looking to design path following controllers for ground vehicles.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1190-1209"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149130735","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

Tackling Snow-Induced Challenges: Safe Autonomous Lane-Keeping With Robust Reinforcement Learning 解决积雪引发的挑战：基于稳健强化学习的安全自主车道保持

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-11 DOI: 10.1109/TIV.2026.3733200

Amin Jalal Aghdasian;Farzaneh Abdollahi;Ali Kamali Eigoli

{"title":"Tackling Snow-Induced Challenges: Safe Autonomous Lane-Keeping With Robust Reinforcement Learning","authors":"Amin Jalal Aghdasian;Farzaneh Abdollahi;Ali Kamali Eigoli","doi":"10.1109/TIV.2026.3733200","DOIUrl":"https://doi.org/10.1109/TIV.2026.3733200","url":null,"abstract":"This paper proposes two new algorithms for the lane keeping system (LKS) in autonomous vehicles (AVs) operating under snowy road conditions. These algorithms use deep reinforcement learning (DRL) to handle uncertainties and slippage. They include Action-Robust Recurrent Deep Deterministic Policy Gradient (AR-RDPG) and end-to-end Action-Robust convolutional neural network Attention Deterministic Policy Gradient (AR-CADPG), two action-robust approaches for decision-making. In the AR-RDPG method, within the perception layer, camera images are first denoised using multi-scale neural networks. Then, the centerline coefficients are extracted by a pre-trained deep convolutional neural network (DCNN). These coefficients, concatenated with the driving characteristics, are used as input to the control layer. The AR-CADPG method presents an end-to-end approach in which a convolutional neural network (CNN) and an attention mechanism are integrated within a DRL framework. Both methods are first trained in the CARLA simulator and validated under various snowy scenarios. Real-world experiments on a Jetson Nano-based autonomous vehicle confirm the feasibility and stability of the learned policies. Over 50 snowy CARLA validation routes, AR-CADPG reduces the lane-tracking RMSE from 0.42 m for DDPG to 0.23 m, corresponding to a 45.2% improvement. It also reduces RMSE by 32.4% compared with AR-DDPG and by 14.8% compared with AR-RDPG. In real-world JetRacer experiments over 15 repeated laps, AR-CADPG achieves an RMSE of 0.043 m, compared with 0.058 m for AR-RDPG, while both controllers complete all laps without lane departure, collision, or manual reset. The AR-CADPG controller also runs at 20.8 FPS on Jetson Nano, confirming real-time embedded feasibility.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1075-1084"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149131072","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

Strategic Co-Design in Formula 1: Balancing Physical Configuration and Race Tactics f1中的战略协同设计：平衡物理配置和比赛战术

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-16 DOI: 10.1109/TIV.2026.3734358

Marc-Philippe Neumann;Giona Fieni;Francesca Furia;Alberto Cerofolini;Vittorio Ravaglioli;Christopher H. Onder;Gioele Zardini

{"title":"Strategic Co-Design in Formula 1: Balancing Physical Configuration and Race Tactics","authors":"Marc-Philippe Neumann;Giona Fieni;Francesca Furia;Alberto Cerofolini;Vittorio Ravaglioli;Christopher H. Onder;Gioele Zardini","doi":"10.1109/TIV.2026.3734358","DOIUrl":"https://doi.org/10.1109/TIV.2026.3734358","url":null,"abstract":"The development of high performance Formula 1 vehicles is a challenging undertaking. Due to the high complexity of the system, the possibilities to increase efficiency and thus enhance performance in many different subsystems are integral part of a successful vehicle. This research presents a practical application and extension of an existing category-theoretic, monotone co-design framework to Formula 1 racing, addressing both physical component selection (aerodynamic configuration, cooling system, battery) and race strategy (energy allocation, pit stops, starting position) across two hierarchically composed abstraction levels. We first analyze lap-specific aspects and show optimal implementations for the choice of crucial components of a hybrid electric powertrain in racing application. Key objectives include minimizing lap time and short-term component wear, by selecting the optimal components and choosing maximum power deployment. Our optimization results demonstrate significant differences in optimal configuration depending on track characteristics. Then, we raise the level of abstraction and extend the framework to a race scenario, where we include strategic aspects of the competition by means of data-based models and optimization results. The goals are minimizing race time and component wear, while considering energy allocation strategies, pit stops and uncertainties that might occur. By means of a case study we show the influence of a successful qualifying session on the race strategy and therefore component wear. This study provides a foundational framework for future advancements in racing strategy, advocating continued innovation and adoption within the racing industry.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1159-1175"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149133874","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

Bidirectional Charging and Computation for Moving and Parked Electric Vehicles in Edge Computing Networks 边缘计算网络中移动与停放电动汽车的双向充电与计算

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-11 DOI: 10.1109/TIV.2026.3733619

Bhargavi Dande;Hung-Yu Wei

{"title":"Bidirectional Charging and Computation for Moving and Parked Electric Vehicles in Edge Computing Networks","authors":"Bhargavi Dande;Hung-Yu Wei","doi":"10.1109/TIV.2026.3733619","DOIUrl":"https://doi.org/10.1109/TIV.2026.3733619","url":null,"abstract":"This paper presents an optimization-driven framework for bidirectional charging and computation resource allocation in electric vehicle (EV) networks. The proposed framework leverages multi-tier time-of-use (TOU) pricing to dynamically regulate energy and computation interactions based on real-time demand and system conditions. Both moving and parked EVs act as energy and computation prosumers, enabling grid-to-vehicle (G2V) and vehicle-to-grid (V2G) energy exchange while simultaneously utilizing or contributing computational resources during both mobility and parking periods. In low-demand intervals, EVs perform cost-efficient charging and computation offloading to edge servers (ESs), whereas in high-demand intervals, they support the system by participating in V2G energy trading and contributing computational resources. This bidirectional interaction enhances resource utilization and improves overall economic efficiency for EV owners. To achieve system-wide cost minimization and revenue maximization, the framework incorporates a mobility-aware route selection mechanism, where EVs select optimal routes and resource interaction strategies based on real-time pricing and resource availability. Simulation results demonstrate that the proposed framework reduces total system cost by up to 27% and increases net profit by 25% on average compared to existing approaches. These results highlight the scalability and economic effectiveness of the proposed framework for future smart grid and intelligent transportation systems (ITS).","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1110-1126"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149131071","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

Effect of Particle Size Distribution on LiDAR-Based 3D Object Detection Under Adverse Weather 粒径分布对恶劣天气下基于lidar的三维目标检测的影响

IF 14.3 1区工程技术

IEEE Transactions on Intelligent Vehicles Pub Date : 2026-10-01 Epub Date: 2026-09-14 DOI: 10.1109/TIV.2026.3733542

Alok Kumar Tiwari;Ajinkya Shinde;Gaurav Sharma;Manisha Pattanaik;Sri Niwas Singh

{"title":"Effect of Particle Size Distribution on LiDAR-Based 3D Object Detection Under Adverse Weather","authors":"Alok Kumar Tiwari;Ajinkya Shinde;Gaurav Sharma;Manisha Pattanaik;Sri Niwas Singh","doi":"10.1109/TIV.2026.3733542","DOIUrl":"https://doi.org/10.1109/TIV.2026.3733542","url":null,"abstract":"LiDAR-based sensors employing optical spectrum signals play a vital role in providing significant information about the target objects in autonomous driving vehicle systems, but adverse weather, such as fog and rain, severely degrades the performance. This work analyzes the role of particle size distributions in LiDAR-based 3D object detection under adverse weather. Mie theory and meteorological optical range (MOR) are utilized to calculate the attenuation and backscattering coefficient. Based on the coefficient values, the point cloud is generated under foggy and rainy conditions. Particularly, under strong and moderate advection fog environments, the fog droplet size follows the Gamma and Junge distributions, whereas a four-parameter generalized Gamma distribution is utilized to articulate the effect of rain. Subsequently, synthetic KITTI and nuScenes datasets based on backscattering coefficient values have been generated, and trained baseline deep neural network models on these synthetic datasets for car, cyclist, and pedestrian cases under different detection difficulties. The implementation results show significant variation in the system's average precision (AP) across changes in target object dimensionality and the nature of adverse weather conditions. From the object's dimensionality perspective under adverse weather, the car class exhibits the highest AP (over 90%), followed by the cyclist class with AP ranging from 77% to 95%, and, due to smaller dimensions and occlusions, the pedestrian class AP is lowest with values ranging from 71% to 85%. Finally, it was observed that Mie's theory-based results closely mimic real-world scenarios and prevent false positives produced by MOR analysis.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"11 10","pages":"1096-1109"},"PeriodicalIF":14.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149131060","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