[ros-users] Navigation with existing localization

Eitan Marder-Eppstein eitan at willowgarage.com
Wed Dec 1 17:28:24 UTC 2010


Peshala,

On Tue, Nov 30, 2010 at 10:33 PM, Peshala Jayasekara
<peshala_24 at yahoo.com>wrote:

> Dear all,
>
> I want to move a Pioneer2Dx using the ROS navigation stack in a desired
> path (eg. say a circle)
> However, we already have an accurate ultrasonic localization system.
>
> Can someone give an idea how I would use the nodes of navigation stack
> (without using the amcl  for laser based localization).
> - which nodes should I be using?
>

You should be able to just swap out amcl for your ultrasonic localization
system as long as your system publishes a transform to tf from the map to
odometric frame. So in the picture here,
http://www.ros.org/wiki/navigation/Tutorials/RobotSetup#Robot_Setup, you'd
just swap out amcl for your node.


> - I think I can avoid obstacles using only the sonar sensors of the robot
>

You can use sonar sensors to avoid obstacles... you can use any sensor you
like as long as you have a way to convert readings to PointCloud messages.
For sonar, this might involve creating a line of points for each sensor
reading corresponding to the cone size at the hit point.


> - I should combine the odometry with the external localization input
>

I'm not sure if I'm interpreting this correctly, but you could use odometry
information to help your external localization system. However, its not
necessary for things to work as long as your localization system publishes
to tf.

Hope this helps,

Eitan


>
> I managed to keyboard-teleop the Pioneer using p2os stack and it works
> fine.
>
> Need some help on this.
>
> Thank in advance
> Peshala
>
>
>
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20101201/000d47c4/attachment-0003.html>


More information about the ros-users mailing list