Automotive, Other
V2X-Radar
cooperative 3D perception
Vehicle–roadside RGB/LiDAR/4D-radar perception collection with Doppler and aligned object boxes.
Access links
Primary places to inspect, download, read about, or reproduce the dataset.
Quick facts
- Vehicle type
- Automotive, Other
- Environment
- Outdoor
- Data origin
- Real-world
- Sensor count
- 8
- Ground truth
- Not reported
- Calibration
- Parameters Only
- Annotations
- Available
Sensor overview
High-level sensor availability before the detailed sensor records below.
Ground truth
Reference scope, method, rate, coverage, and provenance.
Not reported
Calibration and synchronization
Reported calibration level, reproducibility signals, and supporting notes.
- Level
- Parameters Only
- Processed parameters
- Available
- Raw calibration data
- Not reported
- Calibration targets
- Not reported
Parameters only
Processed calibration parameters are reported, but raw calibration data is not reported.
Notes
Official KITTI-layout calib folders and cooperative per-frame YAML metadata are documented. Reprojection/registration residuals are not universal reference accuracy, and complete raw sessions remain unknown.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
Cooperative, vehicle-only and roadside-only products have different units and membership. Conflicting scenario/frame totals remain unknown rather than combined into one aggregate.
Roadside and vehicle sensors form different rigs. Radar files contain sparse Doppler points; raw radar cubes and standalone raw GNSS/IMU or communication logs are not inferred from mounted equipment.
GPS/PTP/PPS clocks plus nearest-frame matching are reported within 20 ms; simulated communication latency is a separate experimental setting.
Five-class object boxes are derived annotations, not independent ego-pose truth. Calibration residuals and swapped task-area axes do not establish a universal spatial error or collection extent.
The official dataset was released in October 2025 and refined in January 2026. Archive calibration files are documented; complete raw calibration sessions remain unverified.
Cameras
Camera 1
- Model
- Basler acA1920-40gc
- Setup
- Multi Camera
- Count
- 3
- Effective cameras
- 3
- Modality
- Rgb
- Resolution
- 1536 x 864
- Rate
- 10 Hz
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Roadside three-camera group; 10 Hz matched sample cadence.
Camera 2
- Model
- Basler acA1920-40gc
- Setup
- Mono
- Count
- 1
- Effective cameras
- 1
- Modality
- Rgb
- Resolution
- 1920 x 1080
- Rate
- 10 Hz
- Shutter
- Not Reported
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Vehicle front camera; 10 Hz matched sample cadence.
IMUs
Not reported
GNSS
Not reported
LiDAR
RoboSense RS-Ruby-80
- Model or name
- RoboSense RS-Ruby-80
- Count
- 1
- Dimensions
- 3
- Rate
- 10 Hz
- Channels
- 80
- Range
- Not reported
RoboSense RS-Ruby-80
- Model or name
- RoboSense RS-Ruby-80
- Count
- 1
- Dimensions
- 3
- Rate
- 10 Hz
- Channels
- 80
- Range
- Not reported
Radar
OCULI EAGLE
- Count
- 1
- Released data
- Point detections / point clouds
- Frame rate
- 10 Hz
- Doppler velocity
- Available
- Elevation angle
- Available
- Configured range
- Not reported
- Field of view
- 112° H × 44° V (per radar)
- Carrier frequency
- 79 GHz
- Notes
- Sparse XYZ/Doppler detections; 10 Hz describes matched dataset samples.
Arbe Phoenix
- Count
- 1
- Released data
- Point detections / point clouds
- Frame rate
- 10 Hz
- Doppler velocity
- Available
- Elevation angle
- Available
- Configured range
- Not reported
- Field of view
- 100° H × 30° V (per radar)
- Carrier frequency
- 77 GHz
- Notes
- Sparse XYZ/Doppler detections; 10 Hz describes matched dataset samples.
Additional sensors
Not reported
Annotations
Reported annotation availability and task support.
- Availability
- Available
- Format
- Five road-user classes with 3D boxes; KITTI-format single-agent and OpenV2V-format cooperative labels.
- Class count
- 5
Semantic segmentation
Not reported
Instance segmentation
Not reported
Object detection
Available
Optical flow
Not reported
Depth ground truth
Not reported
Citation
Citation key and BibTeX kept at the end of the page for reference.
@article{yang2025v2xradar,
title={V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception},
author={Yang, Lei and Zhang, Xinyu and Li, Jun and Wang, Chen and Ma, Jiaqi and Song, Zhiying and Zhao, Tong and Song, Ziying and Wang, Li and Zhou, Mo and Shen, Yang and Lv, Chen},
journal={Advances in Neural Information Processing Systems (NeurIPS)},
year={2025}
}Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy