Dataset associated with submission for handling LiDAR-degenerative environments through LiDAR-radar-inertial fusion with factor graphs.
The datasets consist of sensor measurements (IMU, LiDAR, and FMCW radar) of an aerial robot manually flown through environments challenging for LiDAR sensors. These are a biking tunnel (geometric self-similarity) and a university corridor filled with fog (obscurants). In addition CAD extrinsics between the sensors are provided.
Manual flight along a 8m-wide, 500m-long, straight section of a biking tunnel in Bergen, Norway.
Bag files can be found here
Manual flight in a university environment in the Elektro building of NTNU in Trondheim, Norway. The first corridor flown through after takeoff is filled with dense fog.
Bag files can be found here
Synchronized using microcontroller-based internally developed synchronization/triggering module
The header time stamp for IMU and radar originate from the triggering module, to replay the trigger-stamped LiDAR data from packet topics use this driver.
| Source | Topic | Rate [Hz] |
|---|---|---|
| Triggering | - /sensor_sync_node/trigger_0- /sensor_sync_node/trigger_1 |
- 800 - 10 |
| IMU | /vectornav_node/uncomp_imu |
200 |
| LiDAR | - /os_cloud_node/imu_packets- /os_cloud_node/lidar_packets- /os_cloud_node/metadata |
10 |
| Radar | /radar/cloud |
10 |
All extrinsics are given with respect to the IMU
LiDAR
- translation [x, y, z]
[-0.00171, 0.02149, 0.0358]
- orientation [x, y, z, w]
[0.000462, 0.0008483, 0.0028835, 0.9999954]
Radar
- translation [x, y, z]
[0.07771, 0.02141, -0.03631]
- orientation [x, y, z, w]
[0.953717, 0, -0.3007058, 0]
We thank the Vestland Fylkeskommune for providing access to the Fyllingsdal sykkeltunnel.
