Open greymfm opened 3 years ago
If I am not mistaken, this will depend heavily on the initial pose graph. You could try to enable homotopy to look for different topologies and see if that changes things.
I encountered the same problem. How did you solve it? Looking forward to your reply.
Hello,
First of all : congrats to all of the RST TU Dortmund team - you are doing a great job! :-)
My differential robot (front differential steering) can turn in-place. Sometimes 'teb_local_planner' chooses the most efficient side to turn in-place to the goal pose, sometimes not.
Example: Turning in-place -90 degree would be closer to the goal pose than turning (360-90) degree - but at certain locations (always the same locations) it chooses the less-efficient turning side.
How could I optimize this? The same happens with the differential robot from the 'teb_local_planner' tutorials.
Here's a demo: https://drive.google.com/file/d/1GHitcXlAKp3CfvwvEsnKg1-tFLz5RkJM/view
Thanks for any tips :-) Regards, Alexander