I have a setup where my pose information is greatly delayed ~0.5s from my PointCloud data. Because of this, my TF timestamps lag behind by about the same 0.5s.
When I attempt to visualize this in RVIZ, the queue size is only 10 messages. Since my laser data publishes much faster than that, I drop 100% of message since I would need a queue of ~40 messages to receive enough messages to transform them successfully.
Does RVIZ have a parameter. config, or setting for changing the MessageFilters queue size?
Also, if I use topic tools throttle, I can reduce the rate of incoming messages, prevent the queue from filling up, and then display around 5 to 10 messages per second. This isn't an ideal work around though.
Hmmm, I got luckier with Googling, and I found this response: http://code.ros.org/lurker/message/20100726.173011.b343a0f5.gl.html
And there's the hardcoded value of 10 in PointCloud2 display on line 51. http://www.ros.org/doc/electric/api/rviz/html/point__cloud2__display_8cpp_source.html
Is there a way to change this dynamically or should I recompile RVIZ/make my own plugin?
If there isn't a way to change this, let me know, and I'll file a ticket.
Asked: 2011-10-26 15:54:24 -0500
Seen: 179 times
Last updated: Oct 27 '11
callback issue when using tf::MessageFilter
Updates for tf::MessageFilter example in geometry_tutorials
how does second term of advertise
Do subscribers work on a FIFO queue?
What happens is a node receives a message while it's in the middle of a callback?
Know if you are dropping messages from a subscriber queue?
How to obtain current queue capacity?
Queued messages do not get sent when shutting down a publisher
How to visualize surface normals as Marker::Arrow for each point in rviz?
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.