0
$\begingroup$

Rosanswers logo

Hi, I have had some trouble with gmmaping with turtlebot. What happens is that when I try to make big maps (a whole building of my university) it seems like it divides the map in different sections and, sometimes it moves the sections so the map at the end fits the reality but only by sections.

For example there are 4 main corridors that make a perfect cross. If I start at the very beginning of one corridor at the end I'll have 4 corridors but not necessarily in a 90º each corner.

In order to work in the code I want to record with rosbag the topics that will allow me to simulate that I'm driving the robot. Which ones do you think I should record? I tried with odom, scan and tf and it doesn't work well, although I can see in rviz the scan the robot doesn't move.

Thanks


Originally posted by apalomer on ROS Answers with karma: 318 on 2011-12-07

Post score: 0

$\endgroup$

1 Answer 1

0
$\begingroup$

Rosanswers logo

Gmapping itself only needs the /scan and /tf topics. /odom could also be useful. Gmapping doesn't care about much else. It just needs to know the transforms for base_link->odom and your_laser_scan_frame->base_link. As long as these are being published, gmapping should have no problem. What problems do you encounter when using just these three topics?


Originally posted by DimitriProsser with karma: 11163 on 2011-12-07

This answer was ACCEPTED on the original site

Post score: 0

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.