wrong tof frame id

[tiralonghipol]

after digging and playing around a bit with the Voxl2 (StarlingV2) I think I figured out that the pointcloud coming from the tof sensor is published wrt world frame.
Is this intended? is there any way to change this?

To give some more context, I am using the voxl_mpa_to_ros_node and if I echo the tof topic, here is the output:

voxl2:~$ rostopic echo -n 1 /tof_pc/header
seq: 263
stamp: 
  secs: 1693480645
  nsecs: 400399933
frame_id: "world"

moreover, the tf between the odometry and the fixed world frame is not published (?)
if you can point me where to add this simple code I can do it myself (assuming is accessible)

Hope this is clear, thank you

[tiralonghipol]

@Moderator
any update on this?
Thank you!

[kerct]

@tiralonghipol @Moderator I'm also trying to use the /tof_pc data and would like more information on this world frame. Is the origin wherever the drone was when it boots up? What's the direction of the world frame?

The only information I can find is the transformation between the body frame and the tof camera in the extrinsics file, but I can't find anything about how this relates to the world frame.

[Moderator]

@kerct tof_pc does not relate to world frame. You need something like voxl-mapper to fuse VIO with tof_pc into world frame

[kerct]

@Moderator okay, so I guess the frame ID in the /tof_pc ROS topic is misleading. Is the frame supposed to be the tof camera?

Also just to confirm if my app communicates over MAVROS to obtain the local position, the IMU being used is from imu_px4 and not imu_apps? (I'm using Starling)

[tiralonghipol]

@kerct I ended up building the mpa wrapper from source
There you can modify the frames id and create your own tf tree
took a while but it works now