List of Datasets

UGV, Handheld

M2UD

Uneven-terrain multi-sensor ground-robot SLAM

58 primary ground-robot sequences with lidar, RGB-D, inertial/magnetic and RTK data, challenging weather and motion; reference trajectories use RTK smoothing or SLAM, with scanner reference maps for two scenes.

Metadata: Complete Metadata curator reviewed

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

Quick facts

Vehicle type
UGV, Handheld
Environment
Indoor, Outdoor
Data origin
Real-world
Sensor count
6
Ground truth
Available
Calibration
Partial Raw Data
Annotations
Not reported

Sensor overview

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

Cameras

2 reported

IMUs

1 reported

GNSS

1 reported

LiDAR

1 reported

Magnetometers

1 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 Not reported Not reported Not reported Partial · Outdoor Sequences With Usable RTK. 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.
Hybrid
Accuracy and notes
Accuracy notes
Receiver ±(8 mm+1 ppm) horizontal/±(15 mm+1 ppm) vertical is not achieved smoothed-trajectory accuracy.
Reference notes
RTK positions filtered/smoothed; 5 Hz and lidar-frequency variants are one reference process.
Other Other Not reported Not reported Not reported Partial · Indoor Sequences. 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
Accuracy and notes
Accuracy notes
Endpoint closure 0.031 m after 130 m and 0.1 m after 601 m; not complete absolute trajectory validation.
Reference notes
Offline multi-sensor SLAM reference; exact released orientation scope/rate unestablished.
Other Other Not reported Not reported Not reported Partial · Indoor/Outdoor Mixed Sequences; RTK Factors When Available. 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.
Hybrid
Accuracy and notes
Reference notes
Offline multi-sensor SLAM with available RTK constraints; orientation scope/rate unestablished.
Other Other Reference laser scanner · FARO Focus Premium 70M Not reported Not reported Partial · Park and Dormitory Mapping Reference Scenes. 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.
Hybrid
Accuracy and notes
Accuracy notes
±1 mm ranging and 3.5 mm 3D accuracy at 25 m are scanner specifications, not final registered-map accuracy.
Reference notes
FARO scans registered offline with manual dynamic-ghost removal; reference-only scanner excluded from normal sensor inventory.

Calibration and synchronization

Reported calibration level, reproducibility signals, and supporting notes.

Level
Partial Raw Data
Processed parameters
Available
Raw calibration data
Available
Calibration targets
Self-designed calibration board

Reproducible calibration data

Raw calibration data is reported, so users can reproduce or inspect the calibration process.

Notes

calibration.yaml parameters, board images and a static IMU recording are described. Complete raw reproducibility for every sensor pairing is not established.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

rural_01 has no images; sensor coverage should be checked per sequence.

Primary totals are 58 sequences, 515.6 GB, 41792.6 s and 51008.3 m. Four additional mapping recordings are listed separately and excluded from these totals.

Smoothed RTK and SLAM references lack a uniformly independently validated trajectory error bound.

Darkness, reflections, crowds, smoke, snow, dust and sparse/repetitive geometry challenge SLAM; smoke is artificial.

Converted depth precision and encoding require checking before metric-depth use.

Cameras

Camera 1

Model
Intel RealSense D455
Setup
Rgbd
Count
1
Effective cameras
2
Modality
Rgbd
Resolution
720 x 480
Rate
10 Hz
Shutter
Global
Lens type
Not Reported
HFOV
Not reported
VFOV
Not reported
Notes
One RGB-D system per primary robot setup, color and depth at 10 Hz; rural_01 has no images. Additional mapping recordings are separate from primary totals. Integrated camera IMU is unused.

IMUs

Robot IMU

Model
Xsens MTi 300
Count
1
Accel / gyro
200 / 200 Hz

GNSS

M68UGI-G

Model or name
M68UGI-G
Count
1
Rate
5 Hz
GNSS type
RTK
Position output
Available
Velocity output
Not reported
Heading output
Not reported

LiDAR

Velodyne VLP-16

Model or name
Velodyne VLP-16
Count
1
Dimensions
3
Rate
10 Hz
Channels
Not reported
Range
Not reported

Additional sensors

Magnetometer

Name
Xsens-integrated magnetometer
Count
1
Rate
100 Hz
Notes
Magnetic stream /imu/mag from the same MTi 300, not an additional physical IMU; exact integrated magnetometer model unreported.

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
jia2025m2udv1

BibTeX

@misc{jia2025m2udv1,
  title={M2UD: A Multi-model, Multi-scenario, Uneven-terrain Dataset for Ground Robot with Localization and Mapping Evaluation},
  author={Jia, Yanpeng and Wang, Shiyi and Shao, Shiliang and Wang, Yue and Zhang, Fu and Wang, Ting},
  year={2025},
  eprint={2503.12387},
  archivePrefix={arXiv},
  note={Selected version 1, dated 16 March 2025}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy