0
$\begingroup$

Rosanswers logo

Dear all,

I'm new in the ROS world and I have a question that would seem to be stupid for most of you but we all have to start somewhere.

I have trouble to understand the role of AMCL. I believe that AMCL is used by indoor robots to find a localization of the robot on a pre-loaded map (made with laserscan). Am I wrong?

My robot will be used outdoor with GPS, IMU, Wheel odometer and a laser scan. So I don’t need to include AMCL in the move_base node? Could you confirm that I can use the navigation stack without the AMCL node?

All the transformations from odom to map (publish initially by AMCL) will be managed by localization stack (ukf_localization_node or ekf_localization_node). Am I right?

Best regards


Originally posted by julienH on ROS Answers with karma: 11 on 2015-04-20

Post score: 1


Original comments

Comment by asimay_y on 2016-03-15:
hi, julienH, I wonder to know did you solve this problem? I am planning to do same work as you. can you please give me some instructions? thanks.

Comment by asimay_y on 2016-06-16:
hi, dear @julienH, I wonder to know how did you solve this problem? I am planning to do same work as you. can you please give me some instructions? thanks.

$\endgroup$

1 Answer 1

0
$\begingroup$

Rosanswers logo

AMCL can be used indoors as well as outdoors for localization depending on the application.
So.. the idea is to feed data from GPS, IMU and Wheel Odometer into a kalman filter (extended or unscented) to get a 3D pose estimate of the robot. You can directly use this in your move_base node or you can feed the output pose from the Kalman filter into the AMCL package which will use laser scans, laser based map and will output the (improved) robot's pose estimate. It is recommended to use amcl to improve your pose estimates from the Kalman Filter but you can get away without using AMCL if the output pose estimate from the Kalman filter is accurate. So, if you are satisfied with the output of Kalman filter, you can use navigation stack without AMCL otherwise using AMCL is not hard.

Hope it helps.


Originally posted by Naman with karma: 1464 on 2015-04-20

This answer was ACCEPTED on the original site

Post score: 3


Original comments

Comment by asimay_y on 2016-03-15:
hi, dear, how to feed the output pose from Kalman filter into AMCL package? because as I know, the input of AMCL is based on laser, not pose info. how to do this ? and we also want laser data function well combine with other odom data. Thanks.

Comment by 130s on 2016-06-16:
That should be separately asked as a new question.

$\endgroup$

Your Answer

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