Coordinate systems (frames)

Main frames in the clover package:

• map has its origin at the flight controller initialization point and may be considered stationary. It is shown as a white grid on the image above;
• base_link is rigidly bound to the drone. It is shown by the simplified drone model on the image above;
• body is bound to the drone, but its Z axis points up regardless of the drone's pitch and roll. It is shown by the red, blue and green lines in the illustration;
• navigate_target is bound to the current navigation target (as set by the navigate service);
• terrain is bound to the floor at the current drone position (see the set_altitude service);
• setpoint is current position setpoint;
• main_camera_optical is the coordinate system, linked to the main camera;

Additional frames become available when ArUco positioning system is active:

• aruco_map is bound to the currently active ArUco map;
• aruco_N is bound to the marker with ID=N.

Frames that are bound to the drone are oriented according to the ROS convention: the X axis points forward, Y to the left, and Z up.

3D visualization of the coordinate systems can be viewed using rviz.

tf2

Read more at http://wiki.ros.org/tf2

tf2 ROS package is used extensively in the Clover platform. tf2 is a set of libraries for C++, Python and other programming languages that are used to work with the frames. Internally, ROS nodes publish TransformStamped messages to /tf topic with transforms between frames at certain points in time.

The simple_offboard node can be used to request the drone position in an arbitrary frame by setting the frame_id argument appropriately in a call to get_telemetry service.

tf2 can be used from Python to transform coordinates (for objects like PoseStamped and PointStamped) from one frame to another