Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

AirV2X: Unified Air-Ground Vehicle-to-Everything Collaboration

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read AirV2X unifies vehicles, roadside units, and drones in one large-scale V2X dataset.

desk verdict A genuinely useful V2X dataset with drones, more thorough than most benchmark papers, but the drone sensor geometry needs a coverage analysis before the aerial-perspective claims are fully trusted. read the letter →

arxiv 2506.19283 v3 pith:RUDIUBMF submitted 2025-06-24 cs.CV cs.AIcs.RO

Xiangbo Gao , Yuheng Wu , Fengze Yang , Xuewen Luo , Keshu Wu , Xinghao Chen , Yuping Wang , Chenxi Liu
show 2 more authors
Yang Zhou Zhengzhong Tu
This is my paper
classification cs.CVcs.AIcs.RO
keywords V2Xcollaborativeperceptionvehicle-to-droneUAV-assisteddrivingcooperativedatasetCARLA-AirSimsimulation3DobjectdetectionBEVsemanticsegmentationdronenavigationstrategies
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper sets out to make drone-assisted vehicle-to-everything (V2D) perception a benchmarkable field by introducing AirV2X-Perception, the first large-scale V2X dataset that combines connected vehicles, roadside units, and drones together. The dataset contains 6.73 hours of synchronized camera and LiDAR data from up to 15 agents per scene, collected across urban, suburban, and rural maps, under clear, cloudy, foggy, and rainy weather and day, dusk, and night lighting. If the dataset is accepted, vehicle-to-drone algorithms will have a common training and evaluation ground, and the field can move from scattered simulated scenes to standardized comparisons. The paper's benchmark results suggest drones help most at night and in rural areas, where fixed roadside infrastructure is sparse.

What carries the argument

The load-bearing object is the dataset itself, produced by co-simulating the CARLA driving simulator and the AirSim drone simulator so that ground vehicles, roadside units, and drones operate in one synchronized world with accurate ground-truth labels. Vehicles and roadside units carry a 360-degree 64-channel LiDAR plus surround cameras, while drones carry the same LiDAR with a downward-facing camera, giving aerial bird's-eye views. The dataset encodes three drone navigation strategies—hover, patrol, and escort—so that the effect of drone flight behavior on perception can be measured, and it supplies 3D bounding boxes, segmentation maps, depth maps, and tracking annotations for downstream tasks.

What would settle it

A decisive test would be to mount the dataset's exact sensor configuration on a real drone and vehicle, record a matched set of rural and nighttime scenarios, and fine-tune a vehicle-to-drone fusion model trained on AirV2X on that real data. If the aerial agents stop contributing beyond the vehicle-only baseline, or if the sim-to-real gap dominates the measured differences between fusion methods, the central premise that this simulation is a valid training ground for drone-assisted perception fails.

Watch

Extended reading notes

Core claim

AirV2X-Perception is presented as the first large-scale V2X dataset that covers connected vehicles, roadside units, and drones together, with up to five of each agent type in a single scene. The dataset provides 6.73 hours of synchronized camera and LiDAR data collected by co-simulating CARLA and AirSim across urban, suburban, and rural maps, with clear, cloudy, foggy, and rainy weather and day, dusk, and night lighting. It includes annotations for 3D object detection, BEV semantic segmentation, depth estimation, multi-object tracking, and three drone navigation strategies: hover, patrol, and escort. Benchmarking six collaborative perception algorithms, the paper finds that heterogeneity-aware fusion methods such as HEAL and STAMP achieve the strongest detection results, and that removing drone agents causes the largest accuracy drops in nighttime and rural settings.

Load-bearing premise

The load-bearing premise is that the CARLA-AirSim co-simulation is faithful enough to real drone-assisted driving that algorithms trained on it transfer to real roads, including the usefulness of the drone's downward-facing LiDAR and single downward camera as a bird's-eye perception source.

Editorial extensions

If this is right

  • Vehicle-to-drone algorithms get a shared benchmark: the same splits, sensor geometry, and evaluation metrics can be used across research groups.
  • Fusion architectures can be compared under a consistent heterogeneous setting that includes ground vehicles, fixed infrastructure, and aerial agents.
  • Drone navigation strategy becomes a studyable factor, since hover, patrol, and escort produce measurably different detection and segmentation outcomes.
  • The ablation results imply that fixed-infrastructure-only systems leave accuracy on the table in rural and nighttime scenes, where aerial agents make their largest contribution.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: a practical deployment rule suggested by the results is to use drones as mobile supplements in rural and nighttime corridors while keeping fixed roadside units at dense urban intersections.
  • Beyond the paper: the benchmark assumes lossless communication, so a natural next step is to inject realistic latency and packet loss into the AirV2X data pipeline and measure how collaborative perception degrades.
  • Beyond the paper: if drone trajectories become a controllable input, future systems could learn or plan patrol and escort routes to maximize downstream perception accuracy, an objective the current dataset does not optimize.
  • Beyond the paper: the strongest validation the authors leave implicit is a matched real-world recording with the same drone sensor configuration; sim-to-real transfer remains the open test that would confirm the dataset's usefulness.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper introduces AirV2X-Perception, a simulated vehicle-to-everything (V2X) perception dataset built by co-simulating CARLA and AirSim. The dataset contains 6.73 hours of driving data with up to five connected vehicles, five roadside units (RSUs), and five drones, across urban, suburban, and rural maps, with day, dusk, and night lighting and clear, cloudy, foggy, and rainy weather. Each agent carries camera and LiDAR sensors, with drones using a downward camera and a LiDAR with vertical FOV -30 to -90 degrees. The authors benchmark six collaborative perception methods (When2com, CoBEVT, Where2comm, V2XViT, HEAL, and STAMP) on 3D object detection, BEV semantic segmentation, and GPU memory usage, and provide ablation studies showing the contribution of each agent type and of drone navigation strategies (hover, patrol, escort). The dataset and development kits are released publicly.

Significance. If the dataset is sound, it is a useful contribution: it is among the first to include vehicles, RSUs, and drones together at scale, and the open release with benchmark code and extensive ablations would support standardized evaluation of vehicle-to-drone collaborative perception. The paper also includes machine-checkable benchmark tables and a released repository, which are strengths. However, the central claim of a useful aerial bird's-eye perspective rests on the drone sensor configuration and on the validity of the co-simulated setup, and one of the main comparative analyses appears to contain a severe confounding between drone navigation strategy and lighting condition. These issues need to be resolved before the dataset and its benchmark conclusions can be fully trusted.

major comments (4)
  1. [§4.4, Tables 5 and 6; Tables 4 and 7] The Hover columns in Table 5 are numerically identical to the Night columns in Table 4 for every method (e.g., When2com 15.4/15.0, CoBEVT 10.5/2.0, HEAL 32.8/31.7), and the same exact equality holds between the Hover column of Table 6 and the Night column of Table 7. This is not a rounding coincidence; it indicates that the hover subset is exactly the nighttime subset, or that the tables were generated from identical data. Consequently, the conclusion in §4.4 that hover mode is the most challenging and that patrol mode is best is confounded with lighting conditions: the hover results may simply reflect nighttime performance. The authors must re-run the navigation-strategy analysis with lighting held constant, or explicitly report the intersection of scenario attributes and show that hover and night are not perfectly correlated.
  2. [Table 2, §3.3, §4.2] The drone LiDAR has a vertical FOV of -30 to -90 degrees and the drone camera is downward-facing, yet the paper never reports drone altitudes, the fraction of annotated objects that fall inside any drone's field of view, or an ablation comparing this sensing configuration with a wider or horizon-inclusive FOV. For a detection evaluation area of 281.6 m x 80 m, a drone at 30 m altitude with the stated FOV covers a ground disc of radius about 17 m, so the aerial sensor footprint is a small fraction of the scene. Without coverage statistics or a FOV ablation, the performance gains attributed to 'drone perspective' in Tables 3-5 and Appendix B may be specific to this narrow downward-looking configuration rather than a general property of vehicle-to-drone collaboration. The paper should either provide these analyses or substantially temper the bird's-eye-view claims.
  3. [§3.1, §3.3] The co-simulation of CARLA and AirSim is a load-bearing component of the dataset, but the paper gives no detail on how the two simulators are synchronized in time, how ego-vehicle and drone coordinates are aligned, what latency or drift exists between CARLA actors and AirSim drones, or how sensor streams are timestamped. Since all downstream benchmarks assume synchronized multi-agent inputs, the absence of this information makes it impossible to verify the internal consistency of the dataset. A description of the synchronization protocol, or experiments demonstrating bounded synchronization error, should be added.
  4. [§4.1, Appendix A, Table 3] All benchmark numbers are reported without variance estimates, and Appendix A states that random seeds were fixed across experiments. Claims such as HEAL outperforming STAMP by 1.3 points in AP30 (49.2 vs. 47.9) and HEAL and CoBEVT tying at 33.9 mIoU are therefore not supported by any significance evidence. Multiple seeds with mean and standard deviation, or at least a statement of which differences are stable across seeds, are needed before these ranking conclusions can be considered reliable.
minor comments (5)
  1. [Throughout] There are numerous typographical and spacing errors, including 'UA Vs' in the Abstract, 'CoPercpetion-UA V' and 'UA V3D' in Table 1, and 'experiemental' in §4.5. These should be cleaned up.
  2. [Table A1] The formatting of Table A1 is hard to read: the arrow-and-drop notation (e.g., '16.2 6.8↓') is not explained and the columns are not clearly separated, which obscures the comparison between agent configurations. A clearer layout or a separate delta table is recommended.
  3. [§3.1] The description of which simulator provides the ground vehicles, the map, and the physics, and which provides the drones, is vague. The sentence 'we use it to simulate the dynamics of drone agents and synchronized with actors in the CARLA simulator' is grammatically incomplete and should be rephrased with an explicit architecture description.
  4. [Figure 1 and Figure 3(a)] Figure 1 and Figure 3(a) appear to show the same scene-distribution percentages; consider merging them or labeling them distinctly to avoid redundancy.
  5. [§2.2] The text 'V ehicle + Infrastructure + Dron e' and similar spacing issues in Section 2.1 and 2.2 should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the dataset construction and benchmark are self-contained; author self-citations are incidental and not load-bearing.

full rationale

This is a dataset and benchmark paper, not a derivation. The central claims are that AirV2X-Perception supplies a large-scale, multi-agent, multi-modal simulated V2X dataset with drones, RSUs, and vehicles, and that six published collaboration methods can be evaluated on it. None of these claims is defined in terms of the results it is used to support. Table 1 compares against prior work using attributes that are independently listed; the 'first large-scale unified air-ground V2X dataset' claim is a positioning statement based on those comparisons, not a prediction derived from the dataset itself. The benchmark results in Tables 3-5 and the ablation tables are produced by training published methods on a held-out split with fixed hyperparameters, so the numbers are empirical outputs rather than fitted inputs renamed as predictions. The authors' own STAMP method appears as one of the six baselines, and LangCoop is described in related work, but neither is invoked to justify the validity of the dataset or to force any conclusion; removing those self-citations would not change any stated derivation. The skeptical concerns about the downward-facing drone LiDAR's limited ground footprint and about sim-to-real transfer are legitimate validation and generalizability risks, but they are not instances of circular reasoning: the paper does not claim to derive drone utility from sensor parameters, and it does not use a drone-coverage assumption to define the benchmark outcome. No equation, fitted parameter, uniqueness theorem, or self-citation chain reduces the paper's conclusions to its inputs. Therefore no significant circularity is present.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The dataset paper introduces no free parameters or invented entities. It relies on simulation fidelity, accurate simulator ground truth, and an unexamined sensor configuration for drones.

assumptions (3)
  • domain assumption CARLA and AirSim co-simulation produces sensor data and physics realistic enough for developing and evaluating collaborative perception algorithms.
    Section 3.1 states the dataset is collected by co-simulating CARLA and AirSim; the paper assumes this simulation is a valid proxy for real drone-assisted driving.
  • domain assumption CARLA ground-truth labels (3D boxes, segmentation, depth) are accurate and cover all objects of interest.
    Section 3.4 relies on CARLA simulator to provide accurate ground truth labels for all objects in the scene.
  • ad hoc to paper A downward-pointing LiDAR and camera on drones yields a useful bird's-eye view for collaborative perception.
    Table 2 specifies drone LiDAR vertical FOV -30/-90 and a downward-facing camera, but the paper does not justify this choice or explain how it provides complementary coverage.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AirV2X: Unified Air-Ground Vehicle-to-Everything Collaboration." pith.science (2026). https://pith.science/paper/RUDIUBMF

@misc{pith2026250619283,
  author       = {Pith},
  title        = {Pith review of: AirV2X: Unified Air-Ground Vehicle-to-Everything Collaboration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RUDIUBMF}},
  note         = {Machine review of arXiv:2506.19283}
}
read the original abstract

While multi-vehicular collaborative driving demonstrates clear advantages over single-vehicle autonomy, traditional infrastructure-based V2X systems remain constrained by substantial deployment costs and the creation of "uncovered danger zones" in rural and suburban areas. We present AirV2X-Perception, a large-scale dataset that leverages Unmanned Aerial Vehicles (UAVs) as a flexible alternative or complement to fixed Road-Side Units (RSUs). Drones offer unique advantages over ground-based perception: complementary bird's-eye-views that reduce occlusions, dynamic positioning capabilities that enable hovering, patrolling, and escorting navigation rules, and significantly lower deployment costs compared to fixed infrastructure. Our dataset comprises 6.73 hours of drone-assisted driving scenarios across urban, suburban, and rural environments with varied weather and lighting conditions. The AirV2X-Perception dataset facilitates the development and standardized evaluation of Vehicle-to-Drone (V2D) algorithms, addressing a critical gap in the rapidly expanding field of aerial-assisted autonomous driving systems. The dataset and development kits are open-sourced at https://github.com/taco-group/AirV2X-Perception.

Figures

Figures reproduced from arXiv: 2506.19283 by the authors.

Figure 1
Figure 1. The percentage of each scenario in the dataset including environments, lighting conditions, [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. A conceptual visualization of three different navigation strategies for drones, along with the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. (a) The percentage of each scene in the dataset; (b) LiDAR point clouds visualization for [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Research Challenges and Progress in the End-to-End V2X Cooperative Autonomous Driving Competition

    cs.RO 2025-07 conditional novelty 3.0 of 10

    This paper summarizes the CVPR 2025 V2X cooperative driving challenge, its winning solutions, and the open research problems it reveals.

  2. Automated Vehicles Should be Connected with Natural Language

    cs.MA 2025-06 conditional novelty 3.0 of 10

    A vision paper recommending natural language as the universal communication medium for connected and automated vehicles.

Reference graph

Works this paper leans on

59 extracted references · 41 canonical work pages · cited by 2 Pith papers

  1. [1]

    Perception and sensing for autonomous vehicles under adverse weather conditions: A survey

    Yuxiao Zhang, Alexander Carballo, Hanting Yang, and Kazuya Takeda. Perception and sensing for autonomous vehicles under adverse weather conditions: A survey. ISPRS Journal of Photogrammetry and Remote Sensing , 196:146–177, 2023. 1

    work page 2023

  2. [2]

    Generative ai for autonomous driving: Frontiers and opportunities

    Yuping Wang, Shuo Xing, Cui Can, Renjie Li, Hongyuan Hua, Kexin Tian, Zhaobin Mo, Xiangbo Gao, Keshu Wu, Sulong Zhou, et al. Generative ai for autonomous driving: Frontiers and opportunities. arXiv preprint arXiv:2505.08854, 2025. 1

    arXiv 2025

  3. [3]

    Mambast: A plug-and- play cross-spectral spatial-temporal fuser for efficient pedestrian detection

    Xiangbo Gao, Asiegbu Miracle Kanu-Asiegbu, and Xiaoxiao Du. Mambast: A plug-and- play cross-spectral spatial-temporal fuser for efficient pedestrian detection. arXiv preprint arXiv:2408.01037, 2024. 1

    arXiv 2024

  4. [4]

    Cost analysis of v2i deployment

    Tom Nokes, Ben Baxter, Harry Scammell, Denis Naberezhnykh, and Leonardo Provvedi. Cost analysis of v2i deployment. Technical Report ED13276, Issue No. 5, Ricardo Energy & Environment, August 2020. 2

    work page 2020

  5. [5]

    Where2comm: Communication-efficient collaborative perception via spatial confidence maps

    Yue Hu, Shaoheng Fang, Zixing Lei, Yiqi Zhong, and Siheng Chen. Where2comm: Communication-efficient collaborative perception via spatial confidence maps. Advances in neural information processing systems , 35:4874–4886, 2022. 2, 3, 4, 6, 7, 8, 9

    work page 2022

  6. [6]

    Stamp: Scalable task and model-agnostic collaborative perception

    Xiangbo Gao, Runsheng Xu, Jiachen Li, Ziran Wang, Zhiwen Fan, and Zhengzhong Tu. Stamp: Scalable task and model-agnostic collaborative perception. arXiv preprint arXiv:2501.18616,

    arXiv

  7. [7]

    Langcoop: Collaborative driving with language

    Xiangbo Gao, Yuheng Wu, Rujia Wang, Chenxi Liu, Yang Zhou, and Zhengzhong Tu. Langcoop: Collaborative driving with language. arXiv preprint arXiv:2504.13406, 2025. 2, 4

    arXiv 2025

  8. [8]

    A synthetic benchmark for collaborative 3d semantic occupancy prediction in v2x autonomous driving

    Hanlin Wu, Pengfei Lin, Ehsan Javanmardi, Naren Bao, Bo Qian, Hao Si, and Manabu Tsukada. A synthetic benchmark for collaborative 3d semantic occupancy prediction in v2x autonomous driving. arXiv preprint arXiv:2506.17004, 2025. 2

    arXiv 2025

Show all 59 references
  1. [9]

    Carla: An open urban driving simulator

    Alexey Dosovitskiy, German Ros, Felipe Codevilla, Antonio Lopez, and Vladlen Koltun. Carla: An open urban driving simulator. In Conference on robot learning, pages 1–16. PMLR, 2017. 2, 4

    2017

  2. [10]

    Airsim: High-fidelity visual and physical simulation for autonomous vehicles

    Shital Shah, Debadeepta Dey, Chris Lovett, and Ashish Kapoor. Airsim: High-fidelity visual and physical simulation for autonomous vehicles. In Field and Service Robotics: Results of the 11th International Conference, pages 621–635. Springer, 2018. 2, 4

    2018

  3. [11]

    Opv2v: An open benchmark dataset and fusion pipeline for perception with vehicle-to-vehicle communication

    Runsheng Xu, Hao Xiang, Xin Xia, Xu Han, Jinlong Li, and Jiaqi Ma. Opv2v: An open benchmark dataset and fusion pipeline for perception with vehicle-to-vehicle communication. In 2022 IEEE International Conference on Robotics and Automation (ICRA) , pages 2583–2589,

    2022

  4. [12]

    V2x-sim: Multi-agent collaborative perception dataset and benchmark for autonomous driving

    Yiming Li, Dekun Ma, Ziyan An, Zixun Wang, Yiqi Zhong, Siheng Chen, and Chen Feng. V2x-sim: Multi-agent collaborative perception dataset and benchmark for autonomous driving. IEEE Robotics and Automation Letters , 7(4):10914–10921, 2022. 3

    2022

  5. [13]

    V2x-vit: Vehicle-to-everything cooperative perception with vision transformer

    Runsheng Xu, Hao Xiang, Zhengzhong Tu, Xin Xia, Ming-Hsuan Yang, and Jiaqi Ma. V2x-vit: Vehicle-to-everything cooperative perception with vision transformer. In Computer Vision – ECCV 2022, 2022. 3, 4

    2022

  6. [14]

    Dair-v2x: A large-scale dataset for vehicle- infrastructure cooperative 3d object detection

    Haibao Yu, Yizhen Luo, Mao Shu, Yiyi Huo, Zebang Yang, Yifeng Shi, Zhenglong Guo, Hanyu Li, Xing Hu, Jirui Yuan, and Zaiqing Nie. Dair-v2x: A large-scale dataset for vehicle- infrastructure cooperative 3d object detection. In Proceedings of the IEEE/CVF Conference on Computer ...

    2022

  7. [15]

    V2v4real: A real-world large-scale dataset for vehicle-to-vehicle cooperative perception

    Runsheng Xu, Xin Xia, Jinlong Li, Hanzhao Li, Shuo Zhang, Zhengzhong Tu, Zonglin Meng, Hao Xiang, Xiaoyu Dong, Rui Song, Hongkai Yu, Bolei Zhou, and Jiaqi Ma. V2v4real: A real-world large-scale dataset for vehicle-to-vehicle cooperative perception. In Proceedings of the IEEE/C...

    2023

  8. [16]

    Rcooper: A real-world large-scale dataset for roadside cooperative perception

    Ruiyang Hao, Siqi Fan, Yingru Dai, Zhenlin Zhang, Chenxi Li, Yuntian Wang, Haibao Yu, Wenxian Yang, Jirui Yuan, and Zaiqing Nie. Rcooper: A real-world large-scale dataset for roadside cooperative perception. In Proceedings of the IEEE/CVF Conference on Computer Vision and Patt...

    2024

  9. [17]

    Walter Zimmer, Gerhard Arya Wardana, Suren Sritharan, Xingcheng Zhou, Rui Song, and Alois C. Knoll. Tumtraf v2x cooperative perception dataset. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22668–22677, 2024. 3, 7

    2024

  10. [18]

    Holovic: Large-scale dataset and benchmark for multi-sensor holographic intersection and vehicle-infrastructure cooperative

    Cong Ma, Lei Qiao, Chengkai Zhu, Kai Liu, Zelong Kong, Qing Li, Xueqi Zhou, Yuheng Kan, and Wei Wu. Holovic: Large-scale dataset and benchmark for multi-sensor holographic intersection and vehicle-infrastructure cooperative. In Proceedings of the IEEE/CVF Conference on Compute...

    2024

  11. [19]

    V2x-real: A large-scale dataset for vehicle-to- everything cooperative perception

    Hao Xiang, Zhaoliang Zheng, Xin Xia, Runsheng Xu, Letian Gao, Zewei Zhou, Xu Han, Xinkai Ji, Mingxi Li, Zonglin Meng, Li Jin, Mingyue Lei, Zhaoyang Ma, Zihang He, Haoxuan Ma, Yunshuang Yuan, Yingqian Zhao, and Jiaqi Ma. V2x-real: A large-scale dataset for vehicle-to- everythin...

    2024

  12. [20]

    V2x-r: Cooperative lidar-4d radar fusion for 3d object detection with denoising diffusion

    Xun Huang, Jinlong Wang, Qiming Xia, Siheng Chen, Bisheng Yang, Xin Li, Cheng Wang, and Chenglu Wen. V2x-r: Cooperative lidar-4d radar fusion for 3d object detection with denoising diffusion. arXiv preprint arXiv:2411.08402, 2024. 3

    2024 arXiv

  13. [21]

    Collaboration helps camera overtake lidar in 3d detection

    Yue Hu, Yifan Lu, Runsheng Xu, Weidi Xie, Siheng Chen, and Yanfeng Wang. Collaboration helps camera overtake lidar in 3d detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9243–9252, 2023. 3

    2023

  14. [22]

    Uav3d: A large-scale 3d perception benchmark for unmanned aerial vehicles

    Hui Ye, Rajshekhar Sunderraman, and Shihao Ji. Uav3d: A large-scale 3d perception benchmark for unmanned aerial vehicles. In Advances in Neural Information Processing Systems 37 (NeurIPS 2024) Datasets and Benchmarks Track , 2024. 3

    2024

  15. [23]

    U2udata: A large-scale cooperative perception dataset for swarm uavs autonomous flight

    Tongtong Feng, Xin Wang, Feilin Han, Leping Zhang, and Wenwu Zhu. U2udata: A large-scale cooperative perception dataset for swarm uavs autonomous flight. In Proceedings of the 32nd ACM International Conference on Multimedia (MM ’24) , 2024. 3

    2024

  16. [24]

    Aritra Dutta, Srijan Das, Jacob Nielsen, Rajatsubhra Chakraborty, and Mubarak Shah. Multiview aerial visual recognition (mavrec): Can multi-view improve aerial visual perception? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),

  17. [25]

    Uvcpnet: A uav-vehicle collaborative perception network for 3d object detection

    Yuchao Wang, Peirui Cheng, Pengju Tian, Xiangru Li, Xiaoyu Zhang, and Licheng Jiao. Uvcpnet: A uav-vehicle collaborative perception network for 3d object detection. arXiv preprint arXiv:2406.04647, 2024. 3

    2024 arXiv

  18. [26]

    Griffin: Aerial-ground cooperative detection and tracking dataset and benchmark

    Jiahao Wang, Xiangyu Cao, Jiaru Zhong, Yuner Zhang, Haibao Yu, Lei He, and Shaobing Xu. Griffin: Aerial-ground cooperative detection and tracking dataset and benchmark. arXiv preprint arXiv:2503.06983, 2025. 3

    2025 arXiv

  19. [27]

    Agc-drive: A large-scale dataset for real-world aerial-ground collaboration in driving scenarios

    Yunhao Hou, Bochao Zou, Min Zhang, Ran Chen, Shangdong Yang, Yanmei Zhang, Junbao Zhuo, Siheng Chen, Jiansheng Chen, and Huimin Ma. Agc-drive: A large-scale dataset for real-world aerial-ground collaboration in driving scenarios. arXiv preprint arXiv:2506.16371,

  20. [28]

    Collaborative perception datasets for autonomous driving: A review

    Naibang Wang, Deyong Shang, Yan Gong, Xiaoxi Hu, Ziying Song, Lei Yang, Yuhan Huang, Xiaoyu Wang, and Jianli Lu. Collaborative perception datasets for autonomous driving: A review. arXiv preprint arXiv:2504.12696, 2025. 2

    2025 arXiv

  21. [29]

    Object classification using cnn-based fusion of vision and lidar in autonomous vehicle environment

    Hongbo Gao, Bo Cheng, Jianqiang Wang, Keqiang Li, Jianhui Zhao, and Deyi Li. Object classification using cnn-based fusion of vision and lidar in autonomous vehicle environment. IEEE Transactions on Industrial Informatics , 14(9):4224–4231, 2018. 3 11

    2018

  22. [30]

    Cooper: Cooperative perception for connected autonomous vehicles based on 3d point clouds

    Qi Chen, Sihai Tang, Qing Yang, and Song Fu. Cooper: Cooperative perception for connected autonomous vehicles based on 3d point clouds. In 2019 IEEE 39th International Conference on Distributed Computing Systems (ICDCS) , pages 514–524. IEEE, 2019. 3

    2019

  23. [31]

    Cooperative perception for 3d object detection in driving scenarios using infrastructure sensors

    Eduardo Arnold, Mehrdad Dianati, Robert de Temple, and Saber Fallah. Cooperative perception for 3d object detection in driving scenarios using infrastructure sensors. IEEE Transactions on Intelligent Transportation Systems, 23(3):1852–1864, 2020. 3

    2020

  24. [32]

    Multimodal deep-learning for object recognition combining camera and lidar data

    Gledson Melotti, Cristiano Premebida, and Nuno Gonçalves. Multimodal deep-learning for object recognition combining camera and lidar data. In 2020 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC) , pages 177–182. IEEE, 2020. 3

    2020

  25. [33]

    Depth completion via inductive fusion of planar lidar and monocular camera

    Chen Fu, Chiyu Dong, Christoph Mertz, and John M Dolan. Depth completion via inductive fusion of planar lidar and monocular camera. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 10843–10848. IEEE, 2020. 3

    2020

  26. [34]

    Dsdnet: Deep structured self-driving network

    Wenyuan Zeng, Shenlong Wang, Renjie Liao, Yun Chen, Bin Yang, and Raquel Urtasun. Dsdnet: Deep structured self-driving network. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXI 16 , pages 156–172. Springer, 2020. 3

    2020

  27. [35]

    Vips: Real-time perception fusion for infrastructure-assisted autonomous driving

    Shuyao Shi, Jiahe Cui, Zhehao Jiang, Zhenyu Yan, Guoliang Xing, Jianwei Niu, and Zhenchao Ouyang. Vips: Real-time perception fusion for infrastructure-assisted autonomous driving. In Proceedings of the 28th annual international conference on mobile computing and networking , p...

    2022

  28. [36]

    We need to talk: Identifying and overcoming communication-critical scenarios for self-driving

    Nathaniel Moore Glaser and Zsolt Kira. We need to talk: Identifying and overcoming communication-critical scenarios for self-driving. arXiv preprint arXiv:2305.04352 , 2023. 3

    2023 arXiv

  29. [37]

    V2vnet: Vehicle-to-vehicle communication for joint perception and prediction

    Tsun-Hsuan Wang, Sivabalan Manivasagam, Ming Liang, Bin Yang, Wenyuan Zeng, and Raquel Urtasun. V2vnet: Vehicle-to-vehicle communication for joint perception and prediction. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, P...

    2020

  30. [38]

    When2com: Multi-agent perception via communication graph grouping

    Yen-Cheng Liu, Junjiao Tian, Nathaniel Glaser, and Zsolt Kira. When2com: Multi-agent perception via communication graph grouping. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition , pages 4106–4115, 2020. 3, 4, 6, 7, 8, 9

    2020

  31. [39]

    Coopernaut: End-to-end driving with cooperative perception for networked vehicles

    Jiaxun Cui, Hang Qiu, Dian Chen, Peter Stone, and Yuke Zhu. Coopernaut: End-to-end driving with cooperative perception for networked vehicles. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17252–17262, 2022. 3

    2022

  32. [40]

    V2x-vit: Vehicle-to-everything cooperative perception with vision transformer

    Runsheng Xu, Hao Xiang, Zhengzhong Tu, Xin Xia, Ming-Hsuan Yang, and Jiaqi Ma. V2x-vit: Vehicle-to-everything cooperative perception with vision transformer. In European conference on computer vision, pages 107–124. Springer, 2022. 3, 6, 7, 8, 9

    2022

  33. [41]

    Adaptive feature fusion for cooperative perception using lidar point clouds

    Donghao Qiao and Farhana Zulkernine. Adaptive feature fusion for cooperative perception using lidar point clouds. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 1186–1195, 2023. 4

    2023

  34. [42]

    Learning for vehicle-to-vehicle cooperative perception under lossy communication

    Jinlong Li, Runsheng Xu, Xinyu Liu, Jin Ma, Zicheng Chi, Jiaqi Ma, and Hongkai Yu. Learning for vehicle-to-vehicle cooperative perception under lossy communication. IEEE Transactions on Intelligent V ehicles, 8(4):2650–2660, 2023. 4

    2023

  35. [43]

    Core: Cooper- ative reconstruction for multi-agent perception

    Binglu Wang, Lei Zhang, Zhaozhong Wang, Yongqiang Zhao, and Tianfei Zhou. Core: Cooper- ative reconstruction for multi-agent perception. In 2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 8676–8686. IEEE Computer Society, 2023. 4

    2023

  36. [44]

    Vehicle-infrastructure cooperative 3d object detection via feature flow prediction.arXiv preprint arXiv:2303.10552, 2023

    Haibao Yu, Yingjuan Tang, Enze Xie, Jilei Mao, Jirui Yuan, Ping Luo, and Zaiqing Nie. Vehicle-infrastructure cooperative 3d object detection via feature flow prediction.arXiv preprint arXiv:2303.10552, 2023. 4 12

    2023 arXiv

  37. [45]

    Cocmt: Communication-efficient cross-modal transformer for collaborative perception

    Rujia Wang, Xiangbo Gao, Hao Xiang, Runsheng Xu, and Zhengzhong Tu. Cocmt: Communication-efficient cross-modal transformer for collaborative perception. arXiv preprint arXiv:2503.13504, 2025. 4

    2025 arXiv

  38. [46]

    Senserag: Constructing environmental knowledge bases with proactive querying for llm-based autonomous driving

    Xuewen Luo, Chenxi Liu, Fan Ding, Fengze Yang, Yang Zhou, Junnyong Loo, and Hwa Hui Tew. Senserag: Constructing environmental knowledge bases with proactive querying for llm-based autonomous driving. In Proceedings of the Winter Conference on Applications of Computer Vision, p...

    2025

  39. [47]

    V2x-vlm: End-to-end v2x cooperative autonomous driving through large vision-language models

    Junwei You, Haotian Shi, Zhuoyu Jiang, Zilin Huang, Rui Gan, Keshu Wu, Xi Cheng, Xiaopeng Li, and Bin Ran. V2x-vlm: End-to-end v2x cooperative autonomous driving through large vision-language models. arXiv preprint arXiv:2408.09251, 2024. 4

    2024 arXiv

  40. [48]

    V2x-llm: Enhancing v2x integration and understanding in connected vehicle corridors

    Keshu Wu, Pei Li, Yang Zhou, Rui Gan, Junwei You, Yang Cheng, Jingwen Zhu, Steven T Parker, Bin Ran, David A Noyce, et al. V2x-llm: Enhancing v2x integration and understanding in connected vehicle corridors. arXiv preprint arXiv:2503.02239, 2025. 4

    2025 arXiv

  41. [49]

    V2x-unipool: Unifying multimodal perception and knowledge reasoning for autonomous driving

    Xuewen Luo, Fengze Yang, Fan Ding, Xiangbo Gao, Shuo Xing, Yang Zhou, Zhengzhong Tu, and Chenxi Liu. V2x-unipool: Unifying multimodal perception and knowledge reasoning for autonomous driving. arXiv preprint arXiv:2506.02580, 2025. 4

    2025

  42. [50]

    Cobevt: Cooperative bird’s eye view semantic segmentation with sparse transformers

    Runsheng Xu, Zhengzhong Tu, Hao Xiang, Wei Shao, Bolei Zhou, and Jiaqi Ma. Cobevt: Cooperative bird’s eye view semantic segmentation with sparse transformers. arXiv preprint arXiv:2207.02202, 2022. 4, 6, 7, 8, 9

    2022 arXiv

  43. [51]

    F-cooper: Feature based cooperative perception for autonomous vehicle edge computing system using 3d point clouds

    Qi Chen, Xu Ma, Sihai Tang, Jingda Guo, Qing Yang, and Song Fu. F-cooper: Feature based cooperative perception for autonomous vehicle edge computing system using 3d point clouds. In Proceedings of the 4th ACM/IEEE Symposium on Edge Computing , pages 88–100, 2019. 4

    2019

  44. [52]

    Sicp: Simultaneous individual and cooperative perception for 3d object detection in connected and automated vehicles

    Deyuan Qu, Qi Chen, Tianyu Bai, Hongsheng Lu, Heng Fan, Hao Zhang, Song Fu, and Qing Yang. Sicp: Simultaneous individual and cooperative perception for 3d object detection in connected and automated vehicles. In 2024 IEEE/RSJ International Conference on Intelligent Robots and ...

    2024

  45. [53]

    An extensible framework for open heterogeneous collaborative perception

    Yifan Lu, Yue Hu, Yiqi Zhong, Dequan Wang, Yanfeng Wang, and Siheng Chen. An extensible framework for open heterogeneous collaborative perception. arXiv preprint arXiv:2401.13964,

    arXiv

  46. [54]

    Global drone industry market forecasts: analysts trim their growth predictions again - Unmanned airspace — unmannedairspace.info

    Philip Butterworth Hayes. Global drone industry market forecasts: analysts trim their growth predictions again - Unmanned airspace — unmannedairspace.info. [Accessed 17-05-2025]. 18

    2025

  47. [55]

    Advancements in uav-enabled intelligent transportation systems: A three-layered framework and future directions

    Tanzina Afrin, Nita Yodo, Arup Dey, and Lucy G Aragon. Advancements in uav-enabled intelligent transportation systems: A three-layered framework and future directions. Applied Sciences, 14(20):9455, 2024. 18

    2024

  48. [56]

    Uav-assisted intelligent vehicular networks, 2025

    Dawei Wang and Ruonan Zhang. Uav-assisted intelligent vehicular networks, 2025. 18, 20

    2025

  49. [57]

    V2x communication between connected and automated vehicles (cavs) and unmanned aerial vehicles (uavs)

    Ozgenur Kavas-Torris, Sukru Yaren Gelbal, Mustafa Ridvan Cantas, Bilin Aksun Guvenc, and Levent Guvenc. V2x communication between connected and automated vehicles (cavs) and unmanned aerial vehicles (uavs). Sensors, 22(22):8941, 2022. 19, 20

    2022

  50. [58]

    Modelling and real deployment of c-its by integrating ground vehicles and unmanned aerial vehicles

    Seilendria Ardityarama Hadiwardoyo. Modelling and real deployment of c-its by integrating ground vehicles and unmanned aerial vehicles . PhD thesis, Universitat Politècnica de València,

  51. [59]

    flying RSUs

    Abhishek Gupta and Xavier N Fernando. Latency analysis of drone-assisted c-v2x communica- tions for basic safety and co-operative perception messages. Drones (2504-446X), 8(10), 2024. 19 13 A Experiments Details All experiments were conducted under consistent hardware and soft...

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.