Quadrotor Sensor Data

Sensor data recorded with a quadrotor helicopter around the campus of the Karlsruhe Institute of Technology in challenging outdoor and indoor environments is provided. The following sensors are used:

  • Inertial measurement unit (ADIS 16448)
  • GPS receiver (u-blox LEA-M8T)
  • Air pressure sensor (Freescale MP3H6115A)
  • Magnetometer (Honeywell HMC5843)
  • 2D laser rangefinder (Hokuyo UTM-30LX)

Bagfiles for three trajectories are given:

Date and TimeBagfile NameDescription
October 7, 2019 10:24 a.m.2019-10-07-10-24-03.bagStart open sky; entering between two buildings of 40m height; short indoor section; transition to outdoor.
October 24, 2019 2:51 p.m.2019-10-24-14-51-02.bag2019-10-24-14-51-02.bag Start outdoor; trajectory continues between two buildings; indoor section of almost 2 minutes.
October 24, 2019 3:02 p.m.2019-10-24-15-02-22.bagStart outdoor; difficult GPS conditions when entering between buildings: one is a high-rise building of 56m height; indoor part; end position equals start position.

For each trajectory, two ROS bagfiles are provided: The original files and files with sensor data converted to standard ROS messages. When playing the bagfiles, please make sure the ROS /use_sim_time parameter is set. An exemplary ROS launch file is provided.

Bagfiles with standard ROS messages

These files are denoted by  *_rosStandardFormat.bag. Due to using standard ROS messages, no pseudorange and Doppler measurements are included in these files. However, we included the GPS position and velocity estimations which were calculated using a weighted least squares estimation. For accurate time synchronization we provide the timestamps when GPS or laser measurements were taken as the data of these measurements is available only with delays. In these bagfiles, all topics are related to ROS time. The following messages are included in the bagfiles:

TopicMessage TypeDescription
/imusensor_msgs/ImuIMU data
/fixsensor_msgs/NavSatFixGPS position
/gpsVelocitygeometry_msgs/TwistWithCovarianceStampedGPS velocity
/gpsValidstd_msgs/TimeTimestamp of GPS measurement
/scansensor_msgs/LaserScanLaser rangefinder data
/laserValidstd_msgs/TimeTimestamp of laser measurement
/barometerPressuresensor_msgs/FluidPressureAir pressure data
/magneticFieldsensor_msgs/MagneticField3D magnetic field (body frame)
/navboardTimesensor_msgs/TimeReferenceTime reference relating the sensor board time to ROS time
/tf_statictf2_msgs/TFMessageCoordinate transformations between laser rangefinder, camera, and body frame.

Original bagfiles

The original bagfiles include the sensor data with timestamps generated by a microcrontroller sensor board. Moreover, GPS pseudorange and Doppler measurements are included as well as ephemeris data and Klobuchar parameters. The following messages are included in the bagfiles:

TopicMessage TypeDescription
/scansensor_msgs/LaserScanLaser scanner data
/tf_statictf2_msgs/TFMessageCoordinate transformations between laser rangefinder, camera, and body frame.
/mavlink/frommavros_msgs/MavlinkMavlink messages with sensor data

The Mavlink message definitions are given in the files

  • mavlink_ite.xml
  • gnssRaw.xml
  • common.xml

The first two files are provided here, the common.xml file can be downloaded from mavlink. A brief overview is given in the following table:

Mavlink Message NameDescription
IMU_ITEIMU data (body frame coordinates)
MAG_3D_ITE3D magnetic field (body frame coordinates)
SCALED_PRESSUREAir pressure data
LASER_MEASUREMENT_VALIDTimestamp of laser measurement
GNSS_MEASUREMENT_VALIDTimestamp of GPS measurement
GNSS_RAW_MEASUREMENTSPseudorange and Doppler measurements
GNSS_RAW_IONKlobuchar parameters
GPS_HEALTHGPS satellites health status

When processing pseudorange and Doppler measurements we recommend using the original files.

Coordinate frames

  • Body frame: X: Forward, Y: Right, Z: Down
  • Navigation frame: North, East, Down


The datasets are licensed under a Creative Commons Attibution 4.0 International (CC BY 4.0) license. When using the dataset, please cite our paper.

@INPROCEEDINGS{MuellerAtman2020ITM, author = {Karsten Mueller and Jamal Atman and Nikolai Kronenwett and Gert F. Trommer} title = {A Multi-Sensor Navigation System for Outdoor and Indoor Environments} booktitle = {Proceedings of the 2020 International Technical Meeting of The Institute of Navigation, San Diego, California, January 2020} year = {2020} }

For questions regarding the datasets please contact us (Karsten Müller, karsten.mueller@kit.edu and Jamal Atman, jamal.atman@kit.edu).