Automotive

Oxford Radar RobotCar

Scanning radar odometry, localization and multimodal mapping

Repeated Oxford vehicle traversals combine range-intensity scanning radar, surround sensors and derived planar radar-frame odometry.

Metadata: Complete Metadata source documented Released 2019

Quick facts

Vehicle type
Automotive
Environment
Outdoor
Data origin
Real-world
Sensor count
14
Ground truth
Available
Calibration
Parameters Only
Annotations
Not reported

Sensor overview

High-level sensor availability before the detailed sensor records below.

Ground truth

Reference scope, method, rate, coverage, and provenance.

Ground-truth references for Oxford Radar RobotCar
Scope Method Reference system Count Rate Coverage Reference provenance
External
Measured using equipment independent of the platform’s normal onboard sensor pipeline.
Derived
Computed primarily from the dataset’s own recorded sensors.
Hybrid
Combines independent reference equipment with onboard measurements.
2D + heading Visual SLAM Joint VO/loop-closure/GPS-INS optimization Not reported 4 Hz Full Reference provenance
External
Measured using equipment independent of the platform’s normal onboard sensor pipeline.
Derived
Computed primarily from the dataset’s own recorded sensors.
Hybrid
Combines independent reference equipment with onboard measurements.
Derived
Reference notes
SE(2) relative radar-frame trajectory aligned to scans; source geometry is approximate and depends on onboard visual/navigation observations.

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

SDK includes sensor extrinsics; seed measurements optimized with laser/radar observations.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

Thirty-two January 2019 traversals cover 280 km, including two partial routes. This radar extension has its own release and does not inherit the parent year-long recording totals.

Reference odometry is derived SE(2) motion at radar times; CSV zero height/roll/pitch do not make it 6-DoF truth. Navigation and visual observations are shared with normal inputs.

Radar has no Doppler. Missing azimuths can be interpolated and flagged; binary LiDAR clouds are not motion compensated. Precise storage units and achieved timing/reference errors remain unknown.

Full access requires Google-account authorization; samples are separately available. Extrinsics are provided, while complete raw calibration sessions and long-term rigidity remain unverified.

Cameras

Camera 1

Model
Point Grey Bumblebee XB3 BBX3-13S2C-38
Setup
Trinocular
Count
1
Effective cameras
3
Modality
Rgb
Resolution
1280 x 960
Rate
16 Hz
Shutter
Global
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported

Camera 2

Model
Point Grey Grasshopper2 GS2-FW-14S5C-C
Setup
Fisheye
Count
3
Effective cameras
3
Modality
Rgb
Resolution
1024 x 1024
Rate
11.1 Hz
Shutter
Global
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported

IMUs

NovAtel SPAN-CPT ALIGN

Model
Not reported
Count
1
Accel / gyro
Not reported

GNSS

NovAtel SPAN-CPT ALIGN

Model or name
NovAtel SPAN-CPT ALIGN
Count
1
Rate
Not reported
GNSS type
GPS_IMU
Position output
Available
Velocity output
Not reported
Heading output
Not reported

LiDAR

Velodyne HDL-32E

Model or name
Velodyne HDL-32E
Count
2
Dimensions
3
Rate
20 Hz
Channels
32
Range
Not reported

SICK LMS-151

Model or name
SICK LMS-151
Count
2
Dimensions
2
Rate
50 Hz
Channels
Not reported
Range
Not reported

Radar

Navtech CTS350-X

Count
1
Released data
Scanning power images
Frame rate
4 Hz
Doppler velocity
Not available
Elevation angle
Not available
Configured range
163 m
Field of view
360° H (per radar)
File format
Polar PNG
Notes
No Doppler; 400 azimuths ×3768 range-power bins, configured 163 m.

Additional sensors

Odometry

Name
Released visual odometry
Count
1
Notes
The official directory layout includes vo/vo.csv, distinct from the optimized planar radar reference.

Annotations

Reported annotation availability and task support.

Availability
Not reported
Format
Not reported
Class count
Not reported
Semantic segmentation Not reported
Instance segmentation Not reported
Object detection Not reported
Optical flow Not reported
Depth ground truth Not reported

Citation

Citation key and BibTeX kept at the end of the page for reference.

Citation key
barnes2020radarrobotcar

BibTeX

@inproceedings{barnes2020radarrobotcar,
  address = {Paris},
  author = {Dan Barnes and Matthew Gadd and Paul Murcutt and Paul Newman and Ingmar Posner},
  title = {The Oxford Radar RobotCar Dataset: A Radar Extension to the Oxford RobotCar Dataset},
  booktitle = {Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)}, 
  url = {https://arxiv.org/abs/1909.01300},
  pdf = {https://arxiv.org/pdf/1909.01300.pdf},
  year = {2020}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy