Ask Your Question
0

Is there a way to subscribe to the uncertainty of Gmapping?

asked 2012-07-17 08:22:50 -0600

IFLORbot gravatar image

updated 2012-07-17 08:31:52 -0600

I'm trying to add a bit more localization to the Turtlebot during gmapping via RFID tags, so I have them publishing on the vo for robot_pose_ekf. I have noticed that gmapping runs through the certainty of it's pose in order to determine where it is / read the paper. Is there any way to subscribe to this like it's a topic? I can use this value to train my robot when to trust the RFID tags and when to not, since it's concurrently using them for localization and assigning them positions.

Edit: I'm basically doing SLAM with RFID tags.

edit retag flag offensive close merge delete

Comments

Yeah, I know about the Entropy, but there are a lot more calculations that aren't published. I'll look into it more. Thanks. The Entropy is relatively static in terms of other calculations going on, from what I've observed, so it isn't much good to me on the move-to-move.

IFLORbot gravatar imageIFLORbot ( 2012-07-18 02:53:02 -0600 )edit

1 Answer

Sort by ยป oldest newest most voted
2

answered 2012-07-17 11:05:41 -0600

According to the gmapping wiki page there is a entropy value published which gives some information about the pose uncertainty. Have you tried that?

Other than that, I'm not aware of existing mechanisms. You would probably have to look into that yourself. It shouldn't be too complicated to retrieve the pose estimate for every particle for example, which could then be used to compute a gaussian estimate (e.g. mean and covariance matrix) of the pose array.

edit flag offensive delete link more

Comments

How can you transform this pose uncertainty from a float64 to a matrix which says the covariance???

roberto3 gravatar imageroberto3 ( 2018-02-21 04:19:34 -0600 )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: 2012-07-17 08:22:50 -0600

Seen: 387 times

Last updated: Jul 17 '12