Ask Your Question
0

How does the navigation work in TurtleBot

asked 2011-11-27 18:24:59 -0500

Said gravatar image

updated 2011-11-30 05:52:16 -0500

Murph gravatar image

Hi I would like to ask about the technique that used for Mapping and navigation. dose it build a 3D map and if it how builds it and how uses it for navigation.

Thanks

edit retag flag offensive close merge delete

4 Answers

Sort by ยป oldest newest most voted
2

answered 2011-11-28 07:45:26 -0500

tfoote gravatar image

It uses the ROS navigation stack documented on the ros wiki navigation page

edit flag offensive delete link more

Comments

1
Hi Many thanks for answer But it is most likely 2D navigation , so why TurtleBot uses kinect which is RGB-D scanner?? Cheers
Said gravatar image Said  ( 2011-11-29 14:21:36 -0500 )edit
As mentioned by @mwise_wg and @Stefan Kohlbrecher the 3D data is downsampled to 2D data. Please don't ask follow up questions in "answers" to this question. This is a directed question and answer not a forum. If you have a new follow up question, please pose it as it's own question.
tfoote gravatar image tfoote  ( 2011-12-12 14:06:01 -0500 )edit
1

answered 2011-12-12 11:53:16 -0500

mmwise gravatar image

updated 2011-12-12 11:54:42 -0500

The TurtleBot uses the ROS navigation stack. The navigation stack requires a 2D laser scan to work. To create a 2D laser scan the kinect point cloud data is converted to a laser scan using the pointcloud_to_laserscan package in the turtlebot stack.

The pointcloud_to_laserscan package contains a nodelet called cloud_to_scan which samples the data in a region around specified height in the point cloud. Then given a set of points in an angular slice it returns the min range value for that angular slice. See the code for more details on how a point cloud is converted to a laser scan.

The TurtleBot localization primer might also be instructive: http://www.youtube.com/watch?v=C92bxCr05Gs

edit flag offensive delete link more

Comments

Many Thanks
Said gravatar image Said  ( 2011-12-13 12:55:41 -0500 )edit
0

answered 2011-12-08 12:42:10 -0500

Said gravatar image

Many Thanks

But the 2D navigation methods that you are referring to are they based on line scans on depth images . Any specific "mature" & "robust" method that I can access references about it, also how do they register their RGB-D (Kinect images)

Regards

edit flag offensive delete link more
0

answered 2011-11-29 12:51:01 -0500

Said gravatar image

Hi Many thanks for answer

But it is most likely 2D navigation , so why TurtleBot uses kinect which is RGB-D scanner??

Cheers

edit flag offensive delete link more

Comments

2D navigation methods using laser scans are mature, robust methods that consume modest computational power compared to experimental RGB-D/Visual SLAM methods. The turtlebot also moves in 2D and experiences no attitude changes, so generating a fake 2D laser scan out of the RGB-D data is feasible.
Stefan Kohlbrecher gravatar image Stefan Kohlbrecher  ( 2011-11-29 17:33:31 -0500 )edit
traditional laser scanners are expensive >$500 where as the kinect is ~$100 and can be used to fake a 2D laser scan and provide camera images at the same time. It was an easy choice in the design.
mmwise gravatar image mmwise  ( 2011-11-29 21:25:55 -0500 )edit
Many Thanks But the 2D navigation methods that you are referring to are they based on line scans on depth images . Any specific "mature" & "robust" method that I can access references about it, also how do they register their RGB-D (Kinect images) Regards
Said gravatar image Said  ( 2011-12-11 10:59:29 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

Stats

Asked: 2011-11-27 18:24:59 -0500

Seen: 2,368 times

Last updated: Dec 12 '11