Browse: Home   /   Research   /   Software & Datasets   /   L-CAS Multisensor People Dataset

L-CAS Multisensor People Dataset

Collected by a Pioneer 3-AT robot with a Velodyne VLP-16 3D LiDAR, an ASUS Xtion PRO LIVE RGB-D camera, and a HOKUYO UTM-30LX 2D LiDAR

Description

This dataset contains several rosbag files recorded in one (Minerva Building) of the main buildings of the University of Lincoln. The total length of the data is about 49 minutes. Data are grouped into two classes according to whether the robot was stationary or moving. The robot navigation was controlled with a Logitech F710 Wireless Gamepad using the p2os_driver ROS package. To have an idea of its use, please refer to:

Data format

Sensor data were recorded in their original frame of reference and the transformation between the coordinate frames was implemented by the tf ROS package. The recorded ROS topics and their corresponding specifications are summarised as follows:

Topic (message type) Description Frequency
camera/depth/image raw (sensor msgs/Image) Depth image from the RGB-D camera 30 Hz (18 MB/s)
scan (sensor msgs/LaserScan) Scan data from the 2D LiDAR 40 Hz (120 KB/s)
velodyne packets (velodyne msgs/VelodyneScan) Scan data from the 3D LiDAR 10 Hz (1 MB/s)
pose (nav msgs/Odometry) Odometry data from the robot 10 Hz (7 KB/s)
tf (tf2 msgs/TFMessage) Transform data between coordinate frames 40 Hz (20 KB/s)

 

 

 

 

 

 

 

Citation

If you are considering using this data, please reference the following:

@inproceedings{yz18iros,
   title={Multisensor Online Transfer Learning for 3D LiDAR-based Human Detection with a Mobile Robot},
   author={Zhi Yan and Li Sun and Tom Duckett and Nicola Bellotto},
   booktitle = {In Proceedings of the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
   address = {Madrid, Spain},
   month = {October},
   year = {2018}
}

 

Recording platform

The recording platform includes a Pioneer 3-AT mobile robot with four wheels for indoor use and a built-in odometry sensor, a Velodyne VLP- 16 3D LiDAR (mounted at 80 cm from the floor), an ASUS Xtion PRO LIVE RGB-D camera (55 cm from the floor), and a HOKUYO UTM-30LX 2D LiDAR (11.9 cm from the floor).

Downloads

Class Date Time Data
Stationary 2016/05/23 12:00-12:18 (19 minutes) .rosbag
Moving 2016/05/23 12:27-12:38 (12 minutes) .rosbag 
Stationary 2016/05/23 12:39-12:56 (18 minutes) .rosbag 

License

Creative Commons Licence
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Copyright (c) 2016 Zhi Yan, Tom Duckett, and Nicola Bellotto.

Funding

This work was funded in part by the EU Horizon 2020 project FLOBOT, H2020-ICT-2014-1, Grant agreement no.: 645376.

Acknowledgement

The authors would like to thank Jason Burgon for his help in building aluminium frame and mounting sensors.