Notes Regarding ROS Implementation
The figure below depicts the ROS node graph of the stack when processing camera stream and estimating pose of 2 ESP-drones, espdrone1 and espdrone2.
Transform is also published for each ESP-drone. As the /<drone_name>/aruco_map_pose_tracker node estimates the pose, it is also the node publishing the transform.
The transform published by aruco_map_pose_tracker is from the aruco_map frame to the drone's frame (espdrone1 and espdrone2 in the figure). The pose estimation result in the /<drone_name>/aruco_map_pose_tracker/pose topic is really just the transform from the map frame to the drone's frame.
Both map and aruco_map are fixed frames; the (static) transform from map to aruco_map is for adjusting the coordinate system, e.g. shifting the origin and rotating the axes so that the +z axis points up instead of some direction parallel to the ground. The static transform from map to aruco_map is described in the environment configuration file by the key world_to_marker_map_tf.