ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

Costmap2DROS transform timeout.Could not get robot pose, cancelling reconfiguration

asked 2016-02-26 02:23:07 -0500

YingHua gravatar image

updated 2016-02-26 20:27:51 -0500

Hi , I am doing an experiment about hector slam navigation.

for the picture as below,the small tf mark is where it start and the big one is the last place .

here is my launch file (click me to github)

image description

But the system crash when the laser scanner move every time about 10 seconds.

And shows Costmap2DROS transform timeout and can't get robot pose.

(update stauts 2016.02.27)

I guess this warning is because of the setting I modify.

I modify "hector_navigation/hector_exploration_node/config/costmap.yaml",change " transform_tolerance: 2 " to " transform_tolerance: 10".

image description image description

How could I solve it? Thanks for anyone response.><

Here are the error message.

1000, plan_in_unknown: 1, use_inflated_obstacle: 1, p_goal_angle_penalty_:50 , min_frontier_size: 5, p_dist_for_goal_reached_: 0.250000, same_frontier: 0.250000

tolerance: 10.0000

edit retag flag offensive close merge delete


Your tolerance for transforms is set to 10.000 meaning that the system will not use transform information that is more than 10 seconds old.Your system is not getting any new TF messages after 168.19 . Read about tf and do the tutorials - , then update with more info

vaziri gravatar image vaziri  ( 2016-02-26 16:27:55 -0500 )edit

I make some test yesterday. I guess that SLAM system crash is not because of this warning. Maybe the system sway too much make it crash.But if I want to resolve this warning,should I study more about tf ?

YingHua gravatar image YingHua  ( 2016-02-26 20:19:32 -0500 )edit

Thanks for your response!!vaziri :) I update my question already.

YingHua gravatar image YingHua  ( 2016-02-26 20:28:58 -0500 )edit

did you fix this? im having troubles too

julimen5 gravatar image julimen5  ( 2018-03-24 16:31:14 -0500 )edit

3 Answers

Sort by ยป oldest newest most voted

answered 2019-06-09 05:57:53 -0500

There is a related question (duplicate?) here, will post same answer in both.

In our case it was a performance issue, the computer was not strong enough to manage the load. I changed rviz configuration to prevent pointcloud visualisation, laserscanner vis, and furthermore I was having a process consuming 100% CPU (a node which was not sleeping while spinning). After those fixes, now everything is back to normal.

edit flag offensive delete link more

answered 2020-08-14 03:36:07 -0500

garry97 gravatar image

The problem may be with time synchronization between the machines you are using. See if time on both machines match.

edit flag offensive delete link more


How do you do so ?

tomkimsour gravatar image tomkimsour  ( 2021-05-10 08:53:24 -0500 )edit

answered 2019-12-08 18:07:25 -0500

sritee gravatar image

Sometimes gazebo is running at faster than real time if we wish. Real time factor > 1. In this case, we need to change the map publishing freuquency to a higher number in the navigation stack configuration files.

edit flag offensive delete link more


How would you change it ?

tomkimsour gravatar image tomkimsour  ( 2021-05-10 08:53:35 -0500 )edit

Question Tools



Asked: 2016-02-26 02:23:07 -0500

Seen: 19,691 times

Last updated: Dec 08 '19