Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

From what I understand, the global planner doesn't re-plan unless the local planner fails to execute the current plan, so unless your pose update is large or makes it difficult for the local planner to run, I wouldn't expect navfn to re-plan. That said, if the correction to the pose is small, the local planner can probably handle it without needing to re-plan anyway.

As a side note, sending initial pose messages to AMCL frequently is probably a bad idea, because it will lose any confidence present in the particle filter, and is potentially quite expensive.

On the other hand, if you're only re-localizing infrequently (perhaps when the robot is docked?), then you may just need to wait for AMCL to finish updating its internal state before you publish a goal. Listening on the current_pose topic might be a useful way to get a callback when AMCL is done.