ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

Rviz, slam_gmapping hangs

asked 2013-03-14 03:01:29 -0500

ChengXiang gravatar image

updated 2013-03-15 22:11:22 -0500

Hi, I am having some troubles with Rviz recently, my computer will hangs suddenly while i am using it. Originally it works fine and it is only recently that I got this problem. Does anyone have any idea what is going on? I know that my question sound really vague and lack information, but what I can deduced is my system hangs when I am using Rviz.

My graphic card is GeForce 9300M GS. Driver Version : 304.84. I am using Ubuntu 12.04. ROS fuerte. My Rviz revision number is 1.8.17.

Another thing is, this happened when I am using the 2dnav goal setting function on Rviz. And I think Rviz seems to be unable to receive message from robot footprint and obstacle. I am thinking it might have something to do with cost map or laser/odom input. Just a guess.

Any feedback will be appreciated. Thank you.

Update 1

At the point where it hangs, slam_mapping take up 12% of CPU but used up 1.2 GiB of memory.

After my system recovered from hanging, Rviz takes like 60-70 % of CPU abd used up like 100 plus MIB of memory.And I can no longer find slam_gmapping in the list of process in the system monitor. Yet it appears when I do rosnode list.

After the system recovered from hanging, I can't get any messages from move_base/local_costmap node.

In addition, I got this message on my terminal :

[ WARN] [1363349588.818795860, 121.587000000]: Could not get robot pose,  
cancelling reconfiguration
[ERROR] [1363349591.052922982, 121.887000000]: Extrapolation Error looking up 
robot pose: Unable to lookup transform, cache is empty, when looking up transform 
from frame [/base_footprint] to frame [/map]

A little background on what I am doing, I actually sort of alter the URDF so that the robot's sensors experience some pitching and rolling in motion, so I guess the input to slam_gmapping and costmap is affected. But previously such issue did not happened even after i altered the URDF.

Update 2:

Result from roswtf

Online checks summary:

Found 2 warning(s).
Warnings are things that may be just fine, but are sometimes at fault

WARNING The following node subscriptions are unconnected:
 * /move_base_node/local_costmap/voxel_grid_throttle:
   * /move_base_node/local_costmap/voxel_grid
 * /move_base_node:
   * /move_base_simple/goal
   * /move_base/cancel
 * /voxel_visualizer:
   * /move_base_node/local_costmap/voxel_grid_throttled
 * /gazebo:
   * /gazebo/set_link_state
   * /set_hfov
   * /gazebo/set_model_state
   * /set_update_rate

WARNING These nodes have died:
 * spawn_robot-5
 * slam_gmapping-11


Found 11 error(s).

ERROR Communication with [/robot_state_publisher] raised an error: 
ERROR Communication with [/slam_gmapping] raised an error: 
ERROR Communication with [/point_cloud_to_laser_manager] raised an error: 
ERROR Communication with [/voxel_visualizer] raised an error: 
ERROR Communication with [/base_laser_self_filter] raised an error: 
ERROR Communication with [/rviz_1363420915715223896] raised an error: 
ERROR Communication with [/robot_pose_ekf] raised an error: 
ERROR Could not contact the following nodes:
 * /slam_gmapping

ERROR The following nodes should be connected but aren't:
 * /robot_state_publisher->/robot_pose_ekf (/tf)
 * /robot_pose_ekf->/base_laser_self_filter (/tf)
 * /gazebo->/slam_gmapping (/base_scan/scan)
 * /gazebo->/slam_gmapping (/clock)
 * /robot_state_publisher->/base_laser_self_filter (/tf)
 * /robot_state_publisher->/slam_gmapping (/tf)
 * /slam_gmapping->/robot_pose_ekf (/tf)
 * /robot_pose_ekf->/slam_gmapping (/tf)
 * /robot_pose_ekf->/robot_pose_ekf (/tf)
 * /slam_gmapping->/base_laser_self_filter (/tf)
 * /gazebo- ...
(more)
edit retag flag offensive close merge delete

Comments

Please provide your RViz version and graphics card model. Another thing you could check is if RViz is filling up your memory, which would point to a memory leak.

dgossow gravatar image dgossow  ( 2013-03-14 05:39:22 -0500 )edit

How do I check if Rviz is filling up my RAM?

ChengXiang gravatar image ChengXiang  ( 2013-03-14 05:58:09 -0500 )edit

Recently I got a similar problem several times: Totally unresponsive WLAN and/or RViz, maybe in correlation with kinect pointcloud or navigation's potential map..

felix k gravatar image felix k  ( 2013-03-14 07:14:56 -0500 )edit

You can use Ubuntu's system monitor to check the memory usage of running programs.

dgossow gravatar image dgossow  ( 2013-03-14 13:13:59 -0500 )edit

Please refer to my update 1.

ChengXiang gravatar image ChengXiang  ( 2013-03-15 02:32:23 -0500 )edit

Seems more like slam_gmapping is causing your problem, not rviz. rosnode list might show you nodes that have already crashed, in that case roswtf will give you the right information.

dgossow gravatar image dgossow  ( 2013-03-15 06:42:38 -0500 )edit

Hmmm... I did roswtf and i think i sort of have some idea about what's wrong.

ChengXiang gravatar image ChengXiang  ( 2013-03-15 22:06:21 -0500 )edit

All right. Found the source of my error.

ChengXiang gravatar image ChengXiang  ( 2013-03-15 22:37:41 -0500 )edit

2 Answers

Sort by ยป oldest newest most voted
0

answered 2013-03-15 22:42:31 -0500

ChengXiang gravatar image

Hi peers. Found the source of my error. It is due to me running the robot_pose_ekf package that produce an additional source of /base_link to /odom transformation. This contradict with the transformation produced by Gazebo. After removing the robot_pose_ekf, things went back to normal.

Thanks for all the advices !!!

edit flag offensive delete link more
1

answered 2013-03-14 04:54:50 -0500

TommyP gravatar image

We had a similar problem and the reason was that VSYNC was enabled for the Nvidia graphic. Running nvidia-settings and disabling VSYNC solved the problem.

edit flag offensive delete link more

Comments

Thanks for the reply TommyP. Did you encounter the problem suddenly? For my case, it was originally fine but it just failed to work for no reason suddenly. May I also know when did you encounter this problem?

ChengXiang gravatar image ChengXiang  ( 2013-03-14 05:07:43 -0500 )edit

I did what you mentioned. It doesn't seems to help.

ChengXiang gravatar image ChengXiang  ( 2013-03-14 06:22:18 -0500 )edit

Question Tools

1 follower

Stats

Asked: 2013-03-14 03:01:29 -0500

Seen: 1,092 times

Last updated: Mar 15 '13