Splits odometry into two parts, one of which is resolved via tf.
It allows to rename parent frame.
Converts odometry messages to tf transforms.
Converts tf transforms to odometry messages.
Allows two-phase transform lookup with the root using latest transform without waiting.
This is useful e.g. for fast map-to-base_link odometry from slow map-to-odom transforms.
Converts tf transforms to path messages.
Allows out-of-order transform messages to come and adds older measurements to the correct
place in the path. The individual path points are looked up at times when the stamp_trigger_frame
frame is updated on the tf_msg topic (the idea is that points on the path have the very same timestamps as
your robot's odometry). Do not forget to remap tf_msg to the actual TF topic. stamp_trigger_frame
is by default the same as child_frame, but you have to set it different e.g. if your child_frame
is a static transform to the actual top-level body link.
Use this if a transform you have needs to be republished on a higher frequency
always with current timestamp (to allow constant extrapolation of the transform).
Takes parent_frame->child_frame transform and republishes its inverse as
child_frame->new_parent_frame on publish_frequency, always taking the
latest available transform and stamping it with current time.
Copy a transform to another place in the tree. Reads parent_frame->child_frame
TF and publishes it as new_parent_frame->new_child_frame. If invert_tf is true,
then the transforms gets inverted (just the geometry, not the frame names).
If static_tf is true, the transform is only published when it actually changes,
and is published by a static transform publisher. publish_frequency specifies the
frequency on which the TF is updated, and if it is nonstatic, the also published.
error_if_tf_missing specifies, whether ROS should issue an error in console if the source
TF is missing.
The node filters TF messages by
- including / excluding parent frames,
- including / excluding child frames,
- limiting publishing frequency.
It subscribes topic tf and publishes to tf_filtered.
It allows collecting and republishing transforms in regular intervals.
The tf_connect node allows to connect multiple disconnected TF trees by defining coincident frames.
Some robot drivers incorrectly publish twist in odom frame instead of body frame. This node fixes it.
Some robot drivers do not publish twist or publish a wrong one. This node computes it from differences of the poses.
As above, in nodelets.