List of Datasets

Automotive

NavINST

Multi-sensor autonomous navigation

Vehicle recordings in Canadian urban roads and underground garages, with day/night imagery, multiple IMU grades, two LiDARs, automotive radar and processed pose references.

Metadata: Complete Metadata needs review

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

Download

Not reported

Code

Not reported

Supplementary

Not reported

Quick facts

Vehicle type
Automotive
Environment
Indoor, Outdoor
Data origin
Real-world
Sensor count
22
Ground truth
Available
Calibration
Partial Raw Data
Annotations
Not reported

Sensor overview

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

Cameras

5 reported

IMUs

6 reported

GNSS

2 reported

LiDAR

2 reported

Radar

5 reported

Pitot tube

Not 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.
6DoF GNSS / INS Postprocessed PPK GNSS/IMU reference · Novatel PwrPak7-E1 + KVH1750 Not reported 50 Hz Partial · Outdoor Urban01–Urban06 Trajectories 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 notes
Inertial Explorer reports SD, not independently measured absolute accuracy. Urban01–04 average position SD is around 1 cm and orientation SD below 0.02 degrees; Urban05–06 position SD is around 50 cm, with some-area pitch/roll 0.58 degrees and heading 2.08 degrees.
Reference notes
Geodetic position, ENU velocity and attitude; 1 Hz extended statistics accompany the 50 Hz solution and are not a second reference. ROS/GPS timestamps retained. Reference GNSS/IMU inputs are independently released and thus not independent external truth.
6DoF Other LiDAR-to-map registration with IMU and vehicle odometer Not reported 50 Hz Partial · Indoor Garage Trajectories 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 notes
Published 20 cm RMSE refers to urban validation of the method; comparable indoor performance is expected rather than measured here. Accuracy depends on map quality.
Reference notes
EKF fuses IMU/odometer and periodic LiDAR registration to supplied garage maps, giving local Cartesian 3D position and attitude. Mapping used a stationary LiDAR whose raw measurements/model are not established as a released input stream.

Calibration and synchronization

Reported calibration level, reproducibility signals, and supporting notes.

Level
Partial Raw Data
Processed parameters
Available
Raw calibration data
Available
Calibration targets
Not reported

Reproducible calibration data

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

Notes

Stationary calibration bag, MATLAB camera/stereo projects, OAK–VLP-16 calibration, ROS transformation tree and KVH-origin transforms are supplied. Reference body-to-vehicle alignment is compensated; remaining mechanical distances have stated uncertainty below 1 cm. Full all-sensor calibration reproducibility is not established.

Known or intentional limitations

Reported constraints and characteristics to check before using the dataset.

Outdoor reference uncertainty is substantially worse in Urban05/06 than in Urban01–04.

Indoor reference reuses sensor measurements and its numerical indoor accuracy is not independently demonstrated.

Confined garages and day/night imagery challenge perception methods.

Table III scale and prose release scope require reconciliation; aggregate fields remain unset.

Cameras

oak-mono

Model
Luxonis OAK-1 W PoE
Setup
Mono
Count
1
Effective cameras
1
Modality
Rgb
Resolution
1280 x 720
Rate
30 Hz
Shutter
Global
Lens type
Not Reported
HFOV
127 deg
VFOV
79.5 deg
Notes
Color image and camera-info topics; rounded table megapixel label is not used as an exact pixel count.

zedx-stereo

Model
Stereolabs ZED X
Setup
Stereo
Count
2
Effective cameras
4
Modality
Rgb
Resolution
960 x 540
Rate
15 Hz
Shutter
Global
Lens type
Not Reported
HFOV
110 deg
VFOV
80 deg
Notes
Two physical stereo systems, each with 120 mm baseline and its own integrated IMU. Stereo imagery does not establish a released depth channel.

IMUs

Tactical reference and released IMU

Model
KVH1750
Count
1
Accel / gyro
Not reported

Xsens MTi-670G integrated IMU

Model
Not reported
Count
1
Accel / gyro
Not reported

Livox HAP integrated IMU

Model
Bosch BMI088
Count
1
Accel / gyro
Not reported

ZED X integrated IMUs

Model
Not reported
Count
2
Accel / gyro
Not reported

PwrPak7-E1 integrated IMU, indoor only

Model
EPSON G320N
Count
1
Accel / gyro
Not reported

GNSS

u-blox ZED F9

Model or name
u-blox ZED F9
Count
1
Rate
Not reported
GNSS type
NOT REPORTED
Position output
Available
Velocity output
Not reported
Heading output
Not reported

Novatel PwrPak7-E1

Model or name
Novatel PwrPak7-E1
Count
1
Rate
50 Hz
GNSS type
NOT REPORTED
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
16
Range
Not reported

Livox HAP

Model or name
Livox HAP
Count
1
Dimensions
3
Rate
10 Hz
Channels
Not reported
Range
Not reported

Radar

Smartmicro UMRR-96 Type 153

Count
4
Released data
Point detections / point clouds
Frame rate
20 Hz
Doppler velocity
Available
Elevation angle
Available
Configured range
Not reported
Field of view
Not reported
File format
ROS sensor_msgs/PointCloud2
Notes
Hardware-synchronized 4D ESR point clouds with collective 360-degree coverage. Table specifications list 0.15–19.3 m range and 130/15 degree horizontal/vertical FOV; configured release limits were not independently checked.

OPS241-A

Count
1
Released data
Not reported
Frame rate
20 Hz
Doppler velocity
Available
Elevation angle
Not reported
Configured range
Not reported
Field of view
Not reported
File format
ROS custom message
Notes
Omnipresense front-bumper 1D radar supplies Doppler speed for one ground detection; vehicle forward speed requires mounting-angle calibration/estimation. Table range 1–25 m is a hardware specification.

Pitot tube

Not reported

Additional sensors

Magnetometer

Name
Xsens MTi-670G magnetic-field sensor
Count
1
Notes
/xsens/mag Vector3Stamped is a released topic; chip identity and component rate are not established.

Odometry

Model
FORS-CAN ELM327
Name
Vehicle odometer via OBDII
Count
1
Notes
OBDII automotive scanner gives speed, rpm and maf topics; shared table frequency is 16 Hz. This is not a separate wheel encoder stream.

Other sensor

Name
Xsens MTi-670G pressure stream
Count
1
Notes
/xsens/pressure FluidPressure is listed; the PDF does not explicitly identify a barometric transducer or component sampling rate.

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
dearaujo2025navinst

BibTeX

@misc{dearaujo2025navinst,
  title = {The NavINST Dataset for Multi-Sensor Autonomous Navigation},
  author = {de Araujo, Paulo Ricardo Marques and Mounier, Eslam and Bader, Qamar and Dawson, Emma and Abdelaziz, Shaza I. Kaoud and Zekry, Ahmed and Elhabiby, Mohamed and Noureldin, Aboelmagd},
  year = {2025},
  eprint = {2502.13863},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO}
}

Dataset corrections

Notice a specific issue after reviewing the page?

Report an inaccuracy