Problem with static transform publisher

Question

Hello,

i have a Microsoft Kinect sensor and i am using the freenect driver to use it, Then I transform my pointcloud to a laserscan and show it on my base_link frame. I get the following error:

Transform failure: Lookup would require extrapolation into the past. Requested time 1594648338.669846591 but the earliest data is at time 1594648338.900337345, when looking up transform from frame [kinect_depth_optical_frame] to frame [base_link]

On the Image you can see the current Transformation tree. When you need some information than please tell me.

Regards, markus

---

Originally posted by MarkusHHN on ROS Answers with karma: 54 on 2020-07-13

Post score: 0

Answer 1

Looking at your rqt_tf_tree, I assume the /base2kinect node that is publishing the transform is a static_transform_publisher from the tf library:

http://wiki.ros.org/tf#static_transform_publisher

This uses the tf1 library, which requires you to put a publishing period and uses the /tf topic. Instead, use the tf2 version of the static transform publisher:

http://wiki.ros.org/tf2_ros#static_transform_publisher

This works the same as the tf1 version, but doesn't take a publishing period as the final argument, as it uses the new /tf_static topic which is used for fixed frame transformations. Replacing your tf1 static publisher with this tf2 static publisher should solve your time extrapolation error.

EDIT:

Alternatively, you could add the transform from base_link to kinect_link as a fixed joint in your URDF file.

---

Originally posted by robustify with karma: 956 on 2020-07-13

This answer was ACCEPTED on the original site

Post score: 5

---

Original comments

Comment by Dragonslayer on 2020-07-13:
Is tf2 a ROS2 package, or/and what are the benefits of using it with ROS1?

Comment by MarkusHHN on 2020-07-13:
thank you it works!

Comment by robustify on 2020-07-13:
tf2 is a major improvement and feature enhancement to tf that has been around since ROS Hydro. The tf2 migration guide has the details: http://wiki.ros.org/tf2/Migration. The /tf_static topic is one of the new features, which avoids these timing issues with fixed transforms. In the rqt_tf_tree, when you see a publish rate of 10000 Hz, this means it is a static transform published on /tf_static.