Rviz tf error: unknown reason for transform failure

rviz
tf

asked 2011-06-16 20:05:45 -0600

gong
164 ●9 ●9 ●13

updated 2011-07-03 10:46:22 -0600

mmwise

8372 ●45 ●74 ●99 http://www.meloneewise.com/

Hello everybody! My OS: Ubuntu 10.04

My TF configure: /map -> /odom -> /base_link -> /base_scan. image description

While the TF seems not problem, there is still an fatal error at RVIZ. RVIZ show an error about laser scan: unknown reason for transform failure. I'm confuse, so I need your help. I'm appreciate for any hints.

image description

edit retag flag offensive close merge delete

add a comment

1

answered 2011-06-16 23:23:42 -0600

dornhege

31395 ●134 ●284 ●498

How are you publishing the /scan topic?

It has to have the frame_id /base_scan set. Also it should have a current timestamp (i.e. corresponding to those in the rviz machine), this might be a problem with multiple computers.

Do you see a scan, if you set the fixed_frame in rviz to the frame_id of the scan?

edit flag offensive delete link

Comments

1

when I set /base_scan to be fixed frame in RVIZ, the fatal error is gone.

gong ( 2011-06-16 23:26:50 -0600 )edit

1

In my scan topic, I provide frame_id of scan to be "base_scan". Also, I'm setting the header.stamp of scan message to be current time as timestamp.

gong ( 2011-06-16 23:38:46 -0600 )edit

add a comment

3

answered 2011-09-20 00:51:15 -0600

raahlb
317 ●6 ●13 ●21

I had the same problem; no laser data is shown in rviz unless the laser's frame is set as the fixed frame.

I found a thread on the ros-users list that suggested it had to do with the time_increment field being too large. I noticed there is an error in the LMS1xx package I'm using as a driver for the laser. The fields time_increment and scan_time are ten times as high as they are supposed to be.

I think this was the cause of the error message. What's annoying is that it took 20-30 minutes after I fixed this for it to actually start working. I don't know why it didn't start working immediately. It's possible I restarted rviz at that stage, or managed to reset its laser module. I sure didn't change anything else, at least.

Here's the list thread I mentioned: http://comments.gmane.org/gmane.science.robotics.ros.user/5192

edit flag offensive delete link

Comments

Setting the time_increment to zero solved my problem.

Markus Bader ( 2014-07-22 15:00:41 -0600 )edit

3

For other people trying to make YDLIDAR work -- there's a bug in their ROS package which results in the same issue, and an open PR for it -- https://github.com/EAIBOT/ydlidar/pul....

AlexM ( 2019-07-08 08:50:50 -0600 )edit

add a comment

0

answered 2011-06-16 20:30:13 -0600

mbj
197 ●16 ●18 ●26

I'm not an expert about TF, but did you tried to synchronize the three nodes for publishing at the same frequency? As you can see, the TF /base_link to /base_scan is -0.025 seconds old, maybe the problem is here.

edit flag offensive delete link

Comments

As you said the frequency of the two node is the same. I'm changed the frequency to be different. However, the problem is still there. I will give you the TF at the below.

gong ( 2011-06-16 20:42:22 -0600 )edit

Without any previous change, try to modify the publish frequency of /odom_publisher node to math the average rate of /sick_static_tf_publisher. Maybe this delay affect the next /base_link -> /base_scan TF.

mbj ( 2011-06-16 21:17:04 -0600 )edit

I am seted the odom tf broadcaster to the rate 20.204 Hz as like map tf broadcaster, the problem is the same. Further more, I'm seted the three tf broadcaster to the same, but there have the same error in RVIZ.

gong ( 2011-06-16 21:30:42 -0600 )edit

I'm sorry but this is the only that I know, TF's are not my specialty. :-( Good luck and hope you find the answer soon.

mbj ( 2011-06-16 22:00:58 -0600 )edit

OK, thank you so much!

gong ( 2011-06-16 22:42:31 -0600 )edit

add a comment

0

answered 2011-06-16 20:44:17 -0600