List of Datasets

UGV

M3DGR

sensor-fusion SLAM; sensor degradation

Ground-robot dataset with deliberately induced visual, LiDAR, wheel and GNSS degradation. Current downloads cover a subset; some recordings remain withheld.

Metadata: Complete Metadata source documented Released 2025

Primary places to inspect, download, read about, or reproduce the dataset.

Quick facts

Vehicle type
UGV
Environment
Indoor, Outdoor
Data origin
Real-world
Sensor count
10
Ground truth
Available
Calibration
Parameters Only
Annotations
Not reported

Sensor overview

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

Cameras

3 reported

IMUs

3 reported

GNSS

1 reported

LiDAR

2 reported

Radar

Not reported

Magnetometers

Not reported

Barometers

Not reported

UWB

Not reported

Ground truth

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

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.
3D position RTK GNSS Outdoor RTK · CUAV C-RTK 2HP Not reported 15 Hz Partial 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.
External
Accuracy notes
Advertised 0.8 cm horizontal / 1.5 cm vertical; not scalar trajectory uncertainty.
Reference notes
Outdoor reference groups in Table V; independent device, receiver specifications not dataset error bound.
Other Motion capture Indoor motion capture · Optitrack Not reported 360 Hz Partial 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.
External
Accuracy notes
Table IV lists 1 mm device localization accuracy.
Reference notes
Indoor groups with Mocap; 6DoF mapping requires confirmation because source emphasizes trajectories/position accuracy. Exact orientation/dimensional scope is unestablished; Other does not assert a 6-DoF trajectory.
Other Other ArUco endpoint reference Not reported Not reported Partial 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
Start/end relative transform only; not dense trajectory truth. Camera-derived marker measurements.

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

Maintained calibration.md supplies camera intrinsics, sensor transforms and IMU noise parameters. Complete raw calibration sessions are not established.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

The public subset remains smaller than the thirty-two recorded sequences because long-term recordings are withheld. Collected-corpus duration, distance and storage are not asserted as current public totals.

The paper reports 15 Hz LiDAR versus 10 Hz in maintained release documentation, and 30 Hz RGB-D versus 15 Hz in the same documentation. These rates remain unknown pending edition-specific archive validation.

Indoor motion capture, outdoor RTK and camera-derived ArUco endpoint transforms have different scopes. Endpoint-only references cannot supply dense trajectory ATE, and motion-capture orientation coverage remains unestablished.

Current intrinsic/extrinsic and IMU-noise parameters are published. Raw calibration sessions, native inertial component rates, common sensor triggering and whole-trajectory accuracy remain unknown.

Cameras

Camera 1

Model
RealSense D435i
Setup
Rgbd
Count
1
Effective cameras
2
Modality
Rgbd
Resolution
640 x 480
Rate
Not reported
Shutter
Rolling
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Paper gives 30 Hz RGB/depth, while maintained documentation gives 15 Hz. RGB rolling shutter is explicitly documented; no depth-shutter assumption is made.

Camera 2

Model
Insta360 X4
Setup
Other
Count
1
Effective cameras
1
Modality
Rgb
Resolution
2880 x 1440
Rate
15 Hz
Shutter
Rolling
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
Omnidirectional 360-degree system; physical imager count not stated.

IMUs

D435i integrated 6-axis IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

Livox Avia integrated IMU

Model
BMI088
Count
1
Accel / gyro
Not reported

Livox MID360 integrated IMU

Model
ICM40609
Count
1
Accel / gyro
Not reported

GNSS

CUAV C-RTK 9Ps

Model or name
CUAV C-RTK 9Ps
Count
1
Rate
10 Hz
GNSS type
UNKNOWN
Position output
Not reported
Velocity output
Not reported
Heading output
Not reported

LiDAR

Livox Avia

Model or name
Livox Avia
Count
1
Dimensions
3
Rate
Not reported
Channels
Not reported
Range
Not reported

Livox MID360

Model or name
Livox MID360
Count
1
Dimensions
3
Rate
Not reported
Channels
Not reported
Range
Not reported

Radar

Not reported

Additional sensors

Odometry

Model
WHEELTEC
Name
Wheel odometry
Count
1
Rate
20 Hz
Notes
Differential-drive two wheels; processed odometry is not a stated physical encoder count.

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
zhang2025m3dgr

BibTeX

@misc{zhang2025m3dgr,
  author = {Deteng Zhang and Junjie Zhang and Yan Sun and Tao Li and Hao Yin and Hongzhao Xie and Jie Yin},
  title = {Towards Robust Sensor-Fusion Ground SLAM: A Comprehensive Benchmark and A Resilient Framework},
  year = {2025},
  url = {https://arxiv.org/abs/2507.08364v1}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy