Automotive
KAIST Complex Urban Dataset
Multimodal localization and mapping in complex urban environments
Forty-one automotive sequences with LiDAR and navigation data, including 22 stereo sequences, across 329.109 km of diverse Korean urban routes, with calibration, reconstructed maps, and a qualified SLAM baseline.
Access links
Primary places to inspect, download, read about, or reproduce the dataset.
Quick facts
- Vehicle type
- Automotive
- Environment
- Indoor, Outdoor
- Data origin
- Real-world
- Sensor count
- 12
- Ground truth
- Not available
- Calibration
- Parameters Only
- Annotations
- Not available
Sensor overview
High-level sensor availability before the detailed sensor records below.
Ground truth
Reference scope, method, rate, coverage, and provenance.
Not available
Calibration and synchronization
Reported calibration level, reproducibility signals, and supporting notes.
- Level
- Parameters Only
- Processed parameters
- Available
- Raw calibration data
- Not reported
- Calibration targets
- Checkerboard and structural scenes
Parameters only
Processed calibration parameters are reported, but raw calibration data is not reported.
Notes
Per-sequence encoder, LiDAR, vehicle-frame, and stereo calibration parameters are supplied in Euler and SE(3) forms, including MATLAB and ROS camera files.
Known or intentional limitations
Reported constraints and characteristics to check before using the dataset.
Consumer GPS and VRS-RTK measurements can be sporadic, unreliable, or unavailable around high-rise buildings and other complex structures.
The released 100 Hz graph-SLAM vehicle pose is a baseline, not ground truth; its accuracy varies by sequence and the authors advise against using it as mapping or localization ground truth.
The 22 stereo sequences warn that the reconstructed point cloud should not be assumed to project accurately into the image plane because the SLAM baseline is not sufficiently accurate.
Sensor fields vary by release generation; the earliest sequences provide orientation-only MTi-300 records and older VRS record layouts.
Cameras
Camera 1
- Model
- FLIR FL3-U3-20E4C-C
- Setup
- Stereo
- Count
- 1
- Effective cameras
- 2
- Modality
- Rgb
- Resolution
- 1280 x 560
- Rate
- 10 Hz
- Shutter
- Global
- Lens type
- Not Reported
- HFOV
- Not reported
- VFOV
- Not reported
- Notes
- Two physical cameras release unrectified lossless 8-bit RGGB Bayer PNGs for the 22 LiDAR-navigation-stereo sequences urban18 through urban39; the 19 original LiDAR-navigation sequences do not list stereo data.
IMUs
Consumer-level AHRS
- Model
- Xsens MTi-300
- Count
- 1
- Accel / gyro
- 200 / 200 Hz
GNSS
U-Blox EVK-7P
- Model or name
- U-Blox EVK-7P
- Count
- 1
- Rate
- 10 Hz
- GNSS type
- STANDARD
- Position output
- Available
- Velocity output
- Not reported
- Heading output
- Not reported
SOKKIA GRX 2
- Model or name
- SOKKIA GRX 2
- Count
- 1
- Rate
- 1 Hz
- GNSS type
- RTK
- Position output
- Available
- Velocity output
- Available
- Heading output
- Available
LiDAR
Velodyne VLP-16
- Model or name
- Velodyne VLP-16
- Count
- 2
- Dimensions
- 3
- Rate
- 10 Hz
- Channels
- 16
- Range
- Not reported
SICK LMS-511
- Model or name
- SICK LMS-511
- Count
- 2
- Dimensions
- 2
- Rate
- 100 Hz
- Channels
- Not reported
- Range
- Not reported
Additional sensors
Barometer
- Model
- Withrobot myPressure
- Name
- Relative-altitude sensor
- Count
- 1
- Rate
- 10 Hz
Wheel encoder
- Model
- RLS LM13
- Name
- Left and right wheel encoders
- Count
- 2
- Rate
- 100 Hz
- Notes
- Incremental pulse counts are released.
Other sensor
- Model
- KVH DSP-1760
- Name
- Three-axis fiber-optic gyro
- Variation
- angular_increment
- Count
- 1
- Rate
- 1000 Hz
- Notes
- Releases delta roll, pitch, and yaw between consecutive records.
Annotations
Reported annotation availability and task support.
- Availability
- Not available
- 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.
@article{Jeong2019ComplexUrban, author={Jeong, Jinyong and Cho, Younggun and Shin, Young-Sik and Roh, Hyunchul and Kim, Ayoung}, title={Complex Urban Dataset With Multi-Level Sensors From Highly Diverse Urban Environments}, journal={The International Journal of Robotics Research}, volume={38}, number={6}, pages={642--657}, year={2019}, doi={10.1177/0278364919843996}}Dataset corrections
Notice a specific issue after reviewing the page?
Report an inaccuracy