Skip to main content

Data

Coordinate Frames

Coordinate frames used in this dataset include sensor frames, the body frame, the local frame, the GPS frame and the global frame.

The body frame on the vehicles defined with the X axis pointing forward, the Y axis to the right, and the Z axis pointing down. The origin of the body frame is defined as the center of the rear axle of the vehicle.

Sensors record data relative to an internal sensor frame that then needs to be related to the body frame. The position of a sensor frame relative to the vehicle body frame is described using the extrinsic parameters.

The GPS global frame used within the dataset makes use of the WGS-84 oblate spheroid model of the Earth.

The maps in the dataset are represented in a linearized version of the global frame. Linearized global frames are tied to a latitude and longitude linearization point on the GPS ellipsoid. The dataset follows the North-East-Down convention.

Frames

Sensor Data

Lidars

Each vehicle is equipped with 4 HDL-32E Lidars, we name them - Red, Green, Blue and Yellow. 3D scans from each of these Lidars are provided as VelodyneScan messages from ROS. These messages are converted to PointCloud2 during playback.

sensor data

Cameras

Each log contains images from all 7 cameras from the vehicle. The two front cameras two rear cameras are 1.3 MP stereo pairs operating at 15 Hz. The two side monocular cameras also work at 15 Hz. The 7th front dash camera produces 5 MP images at 6 Hz. All timestamps associated with camera images are times when the cameras were triggered.

All images have been rectified using the intrinsic parameters and stored as PNG images. All 1.3 MP cameras are triggered at the same time such that their images are synchronized with the same timestamp. Each rosbag contains thumbnails from the Front Left camera for visualization.

CENTER

FL
FR
SL
SR
RL
RR

IMU/GPS

The IMU data consists of orientation, linear acceleration and angular velocity. The orientation data represents the orientation of the body frame with respect to local frame (which is where the Imu was started). The angular velocities and linear acceleration values are in rad/s and m/s2 and in the body frame.

The GPS data provides the latitude, longitude and altitude of the body frame with respect to the WGS84 frame

sensor configuration

Calibration

Along with the dataset, we also provide the extrinsic calibration files for all sensors to the body frame. A total of 12 files are provided which includes 7 camera, 4 lidar and 1 imu files. All the calibration values are in the form of yaml files and conform to ROS nomenclature.

Pose

Each log contains the following pose (x, y, z, roll, pitch, yaw) of the vehicle in global frame:

  • Raw Pose – Integrated GNSS pose from Aplannix POS-LV transformed into Global Frame
  • Ground-truth Pose - Pose generated using full bundle adjustment
  • Localized Pose –Pose corrected in (x, y, yaw) using ground reflectivity-based localization
pose

Tools

For easy visualization and development, we provide a set of ROS packages. The packages use ROS tools and libraries. For more details, checkout the Github tutorial here.

tools