Why the SLAM have more than one maps in RVIZ? [closed]
When I'm using gmapping to be SLAM in rviz, with the robot running there will generate more than one maps in RVIZ?
Closed for the following reason
the question is answered, right answer was accepted by
mmwise
close date 2011-06-07 12:18:52
Comments
Could you add the screenshot that you sent to the mailing list as well, since it helps to explain what you mean?
Eric Perko
(
2011-05-17 14:51:01 -0500
)edit
How are you generating/publishing odometry (e.g. p2os or your own custom robot driver)? Is your odometry accurate? If you run the robot around with the fixed frame in rviz set to "/odom" and set the laser scan's "Decay time" to a large number, do you get a mostly consistent looking map?
Eric Perko
(
2011-05-17 14:54:33 -0500
)edit
as you can see from the screenshot, the map is jumping from one place to another
gong
(
2011-05-17 20:08:24 -0500
)edit
I guess I wasn't clear about the "mostly consistent map". This would be without SLAM running, just using your odometry source to place the buffered laser scans into a mostly consistent reference frame.
Eric Perko
(
2011-05-17 20:20:21 -0500
)edit
I set the fixed frames to "/odom" and set the laser scan's "delay time" to 10 seconds. It seems that I do not get the "mostly consistent map", because when the robot is running I cannot see the map in rviz until the robot is stoped.
gong
(
2011-05-17 20:55:55 -0500
)edit
Further more, I got an error message. Error: m_distanceThesholdCheck overridden!!! ... the odometry has a big jump here. this is probably a bug in the odometry/laser input. This error has been happened someday ago, I rewrite the odomety program to fix it. At that time, the formula was wrong.
gong
(
2011-05-17 21:00:43 -0500
)edit
Additional, I published odometry using the relative distance computed from odometry which provided by driver. The odmetry was like that : x = (current_odometry - last_odometry)*cos(theta), y = (current_odometry - last_odometry)*sin(theta), the theta was absolute value provided by gyro.
gong
(
2011-05-17 21:07:09 -0500
)editadd a comment
