IEEE Transactions on Intelligent Vehicles最新文献

{"title":"Perception and Planning of Intelligent Vehicles Based on BEV in Extreme Off-Road Scenarios","authors":"Jingjing Fan;Lili Fan;Qinghua Ni;Junhao Wang;Yi Liu;Ren Li;Yutong Wang;Sanjin Wang","doi":"10.1109/TIV.2024.3392753","DOIUrl":"https://doi.org/10.1109/TIV.2024.3392753","url":null,"abstract":"In extreme off-road scenarios, autonomous driving technology holds strategic significance for enhancing emergency rescue capabilities, reducing labor intensity, and mitigating safety risks. Challenges such as adverse conditions, complex terrains, unstable satellite signals, and lack of roads pose serious safety challenges for autonomous driving. This perspective first delves into a Bird's Eye View (BEV)-based perception-planning framework, aiming to enhance the adaptability of intelligent vehicles to their environment. Subsequently, this perspective further discusses key issues such as Cyber-Physical-Social Systems (CPSS), foundation models, and technologies like Sora for off-road scenario generation, vehicle cognitive enhancement, and autonomous decision-making. Ultimately, the discussed framework is poised to endow intelligent vehicles with the capability to perform challenging tasks in complex off-road scenarios, realizing a more efficient, safer, and sustainable transportation system, thereby providing better support for the low-altitude economy.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 4","pages":"4568-4572"},"PeriodicalIF":8.2,"publicationDate":"2024-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141315236","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":"VRSTNN: Visual-Relational Spatio-Temporal Neural Network for Early Hazardous Event Detection in Automated Driving Systems","authors":"Dannier Xiao;Mehrdad Dianati;Paul Jennings;Roger Woodman","doi":"10.1109/TIV.2024.3392589","DOIUrl":"https://doi.org/10.1109/TIV.2024.3392589","url":null,"abstract":"Reliable and early detection of hazardous events is vital for the safe deployment of automated driving systems. Yet, it remains challenging as road environments can be highly complex and dynamic. State-of-the-art solutions utilise neural networks to learn visual features and temporal patterns from collision videos. However, in this paper, we show how visual features alone may not provide the essential context needed to detect early warning patterns. To address these limitations, we first propose an input encoding that captures the context of the scene. This is achieved by formulating a scene as a graph to provide a framework to represent the arrangement, relationships and behaviours of each road user. We then process the graphs using graph neural networks to identify scene context from: 1) the collective behaviour of nearby road users based on their relationships and 2) local node features that describe individual behaviour. We then propose a novel visual-relational spatio-temporal neural network (VRSTNN) that leverages multi-modal processing to understand scene context and fuse it with the visual characteristics of the scene for more reliable and early hazard detection. Our results show that our VRSTNN outperforms state-of-the-art models in terms of accuracy, F1 and false negative rate on a real and synthetic benchmark dataset: DOTA and GTAC.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"7016-7029"},"PeriodicalIF":14.0,"publicationDate":"2024-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144500938","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":"Advance-FL: A3C-Based Adaptive Asynchronous Online Federated Learning for Vehicular Edge Cloud Computing Networks","authors":"Guifu Ma;Yougang Bian;Hongmao Qin;Chenlong Yin;Chaoyi Chen;Shengbo Eben Li;Keqiang Li","doi":"10.1109/TIV.2024.3392339","DOIUrl":"https://doi.org/10.1109/TIV.2024.3392339","url":null,"abstract":"Enhancing autonomous driving through Federated Learning (FL) in Intelligent Connected Vehicles (ICVs) confronts challenges like limited scalability of central management, computational strains on diverse ICVs, and inefficiencies due to stragglers. This paper presents <inline-formula><tex-math>$mathit{Advancehbox{-}FL}$</tex-math></inline-formula>, a deep reinforcement learning based <inline-formula><tex-math>$underline{mathit{A}}$</tex-math></inline-formula><inline-formula><tex-math>$underline{mathit{d}}$</tex-math></inline-formula>apti<inline-formula><tex-math>$underline{mathit{v}}$</tex-math></inline-formula>e <inline-formula><tex-math>$underline{mathit{a}}$</tex-math></inline-formula>sy<inline-formula><tex-math>$underline{mathit{nc}}$</tex-math></inline-formula>hronous onlin<inline-formula><tex-math>$underline{mathit{e}}$</tex-math></inline-formula> <inline-formula><tex-math>$underline{mathit{F}}$</tex-math></inline-formula>ederated <inline-formula><tex-math>$underline{mathit{L}}$</tex-math></inline-formula>earning with a computation offloading assisted framework for vehicular edge cloud computing networks to mitigate the above challenges. Innovatively, <inline-formula><tex-math>$mathit{Advancehbox{-}FL}$</tex-math></inline-formula> incorporates the concept of “straggler rate”, a metric originally introduced in this study to quantify the degree of lag in participant computation and training, thus enabling targeted mitigation strategies. By employing an asynchronous advantage actor-critic approach for adaptive data offloading and dynamic local iteration adjustments, <inline-formula><tex-math>$mathit{Advancehbox{-}FL}$</tex-math></inline-formula> effectively alleviates computation resource shortages and harmonizes the balance between model accuracy and straggler impact by dynamically managing the straggler rate. Critical findings include over 62% reduction in training times and a 2%<inline-formula><tex-math>$sim$</tex-math></inline-formula>9% increase in model accuracy across varying non-IID data scenarios and reducing training time by more than 6 times as the number of ICVs increases, compared to prevailing methods. Additionally, experiments in both static and dynamic test-bed further validate <inline-formula><tex-math>$mathit{Advancehbox{-}FL}$</tex-math></inline-formula>’s superior scalability and robustness over state-of-the-art approaches, particularly in maintaining high performance under straggler effects, and showcasing robust adaptability across long-term operations, large-scale datasets and abnormal situations.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6971-6989"},"PeriodicalIF":14.0,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144500941","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":"Progressive Contextual Trajectory Prediction With Adaptive Gating and Fuzzy Logic Integration","authors":"Neha Sharma;Chhavi Dhiman;S. Indu","doi":"10.1109/TIV.2024.3391898","DOIUrl":"https://doi.org/10.1109/TIV.2024.3391898","url":null,"abstract":"Despite the rapid advancement of highly automated vehicles poised to mitigate accidents caused by human errors, understanding the behaviors of road users, especially vulnerable pedestrians, remains a significant challenge. The evolution of pedestrian trajectory prediction, transitioning from early motion models to recent deep learning approaches, has highlighted persistent challenges in accurately predicting future trajectories, particularly in complex scenarios. To address this, this paper presents a Progressive Contextual Trajectory Prediction with Adaptive Gating and Fuzzy Logic Integration (PCTP-AGFL). The proposed method incorporates a dynamic progressive generator (DPG) comprising multiple LSTM layers that adapt progressively to pedestrian motion pattern complexities. The DPG is trained using a learned scheduled sampling strategy implemented through an Adaptive Gating Mechanism (AGM), allowing dynamic switching between teacher forcing and normal mode. This is augmented with an Encoder-Decoder Contextual Attention (EDCA) module to enhance contextual awareness. Additionally, a novel Adaptive Fuzzified Discriminator (AFD) is introduced to enhance the model's capability to handle ambiguous trajectories. Experimental results on JAAD/PIE and ETH/UCY datasets demonstrate the method's superiority over baselines and state-of-the-art approaches. Furthermore, a comprehensive ablation study is carried out to tune the progression parameters, training strategy, and the type of classifier logic in the discriminator.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6960-6970"},"PeriodicalIF":14.0,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144501947","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":"Adaptive Entropy Multi-Modal Fusion for Nighttime Lane Segmentation","authors":"Xinyu Zhang;Xiangchen Yin;Xin Gao;Tianheng Qiu;Li Wang;Guizhen Yu;Yunpeng Wang;Guoying Zhang;Jun Li","doi":"10.1109/TIV.2024.3392413","DOIUrl":"https://doi.org/10.1109/TIV.2024.3392413","url":null,"abstract":"Lane segmentation at night is a challenging problem in autonomous driving perception, which is beneficial to improve the robustness of the application. Existing methods has shown great performance in the benchmark dataset, however, they do not consider the bad lighting scenes in practical applications, for example, the performance of the lane segmentation algorithm will be greatly affected at night. In this paper, we propose a novel multi-modal nighttime lane segmentation algorithm, which utilizes camera and LiDAR for complementary information. We illustrate the role of image entropy in showing the distribution of light at night, and propose an adaptive entropy fusion method to obtain the spatial relationship between entropy and modalities to adapt to different lighting scenes. The features of narrow and long lanes are more likely to be lost at night, a lane feature enhancement module is proposed to enhance the network's ability to capture lane features. Extensive experiments and analysis demostrate the effectiveness of our method against the state-of-the-art semantic segmentation and lane segmentation approaches on SHIFT dataset at night. Extensive experiments conducted on SHIFT dataset at night demonstrate that the proposed method achieves the state-of-the-art performance, 88.36%@14.06fps and 87.24%@26.88fps on SHIFT dataset at night, having the capability for real-time applications.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6990-7002"},"PeriodicalIF":14.0,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144500936","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":"Distributed Secure Surrounding Control for Multiple USVs Against Deception Attacks: A Stackelberg Game Approach With Reinforcement Learning","authors":"Qiang Wang;Chun Liu;Jianglin Lan;Xiaoqiang Ren;Yizhen Meng;Xiaofan Wang","doi":"10.1109/TIV.2024.3392423","DOIUrl":"https://doi.org/10.1109/TIV.2024.3392423","url":null,"abstract":"This article investigates the Stackelberg game-based distributed secure surrounding control (SSC) problem for multiple unmanned surface vehicles (USVs) with unknown dynamics under deception attacks (DAs). The proposed scheme is rooted in the Stackelberg game with reinforcement learning (RL), in which the attacker and controller respectively play the roles of the follower and leader, making sequential decisions of the DA and the SSC. Specifically, a distributed target estimator is established to access the target position. By utilizing this estimated position to formulate an intermediate control law, the target surrounding scenario is effectively transformed into the Stackelberg game-solving problem. The RL approach with a neural networks-based actor-critic learning structure is deployed to directly derive the distributed optimal SSC and DA policies from the Bellman error, whilst learning the unknown dynamics of USVs. Moreover, a value function decomposition technique is applied optimally for using the designed control parameters, thereby accelerating the acquisition of optimal policies. A rigorous theoretical analysis is employed to ensure the closed-loop stability of multiple USVs. Simulations are provided to validate the effectiveness of the proposed distributed SSC scheme for multiple USVs.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"7003-7015"},"PeriodicalIF":14.0,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144500940","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":"Event-Triggered Position Scheduling Based Platooning Control Design for Automated Vehicles","authors":"Palanisamy Selvaraj;Ramalingam Sakthivel;Oh-Min Kwon;Rathinasamy Sakthivel","doi":"10.1109/TIV.2024.3391302","DOIUrl":"https://doi.org/10.1109/TIV.2024.3391302","url":null,"abstract":"This paper focuses on the design and implementation of a sampled-data controller for connected autonomous vehicle platoons operating in a predecessor-follower configuration. Due to the cost and reliability concerns associated with velocity and acceleration sensors, this study involves the development of an event-based sampled-data controller that relies solely on position measurements. Considering the limitations of velocity and acceleration sensors, a memory-based sampled-data controller is proposed that utilizes current and preceding position data information to approximate velocity and acceleration. To conserve communication resources, the controller incorporates a dynamic event-driven communication mechanism. In particular, event-driven communication thresholds are adaptively adjusted based on platooning errors between vehicles. This enhances resource utilization while maintaining control performance. In addition, determining the maximum allowable sampling period and event-triggered constraint parameters is crucial for reliable control performance. This is achieved by formulating and solving stability criteria for the closed-loop platoon error system using Lyapunov stability theory and the linear matrix inequality framework. Finally, comprehensive numerical simulations demonstrate the effectiveness of the proposed event-triggered control algorithm under the influence of some key factors, including triggering instants and unknown nonlinearity effects.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6926-6935"},"PeriodicalIF":14.0,"publicationDate":"2024-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144501033","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":"Human–Machine Shared Control for Industrial Vehicles: A Personalized Driver Behavior Recognition and Authority Allocation Scheme","authors":"Yang Zhang;Jianwei Lu;Guang Xia;Amir Khajepour","doi":"10.1109/TIV.2024.3389952","DOIUrl":"https://doi.org/10.1109/TIV.2024.3389952","url":null,"abstract":"Human-machine shared control has become a transitional solution from manual driving to autonomous driving, attracting increasing attention. In order to effectively enhance the stability of the vehicle and achieve adaptive collaborative control between the machine and human driver with different operating behaviors, this paper proposes a novel personalized steering behavior recognition method. On this basis, an adaptive authority allocation scheme is designed to alleviate the workload of the drivers and improve vehicle stability. Firstly, to identify different driving styles accurately, the K-means algorithm is modified based on the adaptive variation of weight coefficients. Subsequently, the numerical functions, considering tracking accuracy and steering error, are utilized to individually develop the reference models for various steering behaviors. Further optimal allocation problems involving driving workload, tracking performance, authority smoothness, and stability are addressed through MPC. Simulation results indicate that the proposed driver recognition method can distinguish ambiguous driving behaviors, and the shared steering control strategy demonstrates superior performance in path tracking, stability enhancing, and driver workload alleviating.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6869-6880"},"PeriodicalIF":14.0,"publicationDate":"2024-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144502818","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":"Agile Flights Through a Moving Narrow Gap for Quadrotors Using Adaptive Curriculum Learning","authors":"Mengyun Wang;Shengde Jia;Yifeng Niu;Yunzhuo Liu;Chao Yan;Chang Wang","doi":"10.1109/TIV.2024.3391384","DOIUrl":"https://doi.org/10.1109/TIV.2024.3391384","url":null,"abstract":"Fast and agile flying through a gap is challenging for a quadrotor if the gap is narrow, tilted, and moving. Due to the strict time-variant position and attitude constraints, collision-free traversal trajectories under the under-actuated quadrotor dynamics are sparse and difficult to solve. To achieve this challenging task in the real world, we propose a Gap-Traversing Adaptive Curriculum Learning (GTACL) approach, which consists of adaptive curriculum reinforcement learning (ACRL) and online thrust updating (OTU). First, ACRL is introduced to improve sample efficiency, and the policy training is accelerated by designing a curriculum adapted to the agent's capability. Second, OTU is proposed to map the acceleration commands to low-level throttle signals by estimating the thrust model during flight, which reduces the intermediate control variables and helps sim2real transfer. We use the prioritized experience replay mechanism that considers both policy update contribution and data acquisition time to adapt to the changing tasks. GTACL is trained entirely in simulation and can be transferred to other quadrotors with different dynamics. Furthermore, we achieve zero-shot transfer to the real-world quadrotor without fine-tuning. The average success rates of 98% and 87.8% in simulation and real-world experiments for different task conditions demonstrate the robustness of the proposed approach. Comparative results with traditional and related learning-based approaches show the advantages of GTACL in terms of learning efficiency, control performance, and generalization.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6936-6949"},"PeriodicalIF":14.0,"publicationDate":"2024-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144501939","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":"Distributed Adaptive Cooperative Control for Human-in-the-Loop Heterogeneous UAV-UGV Systems With Prescribed Performance","authors":"Hongjing Liang;Shoufeng Yang;Tieshan Li;Huaguang Zhang","doi":"10.1109/TIV.2024.3391176","DOIUrl":"https://doi.org/10.1109/TIV.2024.3391176","url":null,"abstract":"This paper focuses on the distributed adaptive cooperative control problem for human-in-the-loop (HiTL) heterogeneous unmanned aerial vehicle-unmanned ground vehicle (UAV-UGV) systems via an improved prescribed performance approach. A novel human motion recognition system (HMRS) is designed and integrated into the HiTL strategy. Specifically, the leader trajectory can be changed in real time based on HMRS and the leader signal library to cope with various unexpected situations. This HiTL strategy solves the discontinuity and non-differentiability problems that may exist before and after human modification of leader signals within the extremely short time in conventional HiTL strategies. Moreover, an improved predefined-time prescribed performance approach is proposed, in which the performance function only needs to be first-order continuous differentiable instead of infinite-order continuous differentiable. This approach can greatly broaden the selection range of performance functions. Furthermore, a unified model of heterogeneous UAV-UGV systems is established to avoid designing UAV and UGV systems separately, which improves the universality of the control algorithm. Finally, the proposed HiTL control scheme is applied to a simulation example to verify its feasibility and effectiveness.","PeriodicalId":36532,"journal":{"name":"IEEE Transactions on Intelligent Vehicles","volume":"9 11","pages":"6912-6925"},"PeriodicalIF":14.0,"publicationDate":"2024-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144500935","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}