Hi @staff, happy new year!
To begin this year I have the following doubt:
I am working with path planners in Navigation Stack and applied DWA, EBand,TEB and MPC to my car-like robot. Unfortunately just TEB works good as local planner for this robot type…
For this reason I wish to try new path planning algorithms that I have found on internet, such as those of OMPL and the ones built properly to non-holonomic robot (car, ackermann…) described here:
- Mpc Applied, TESTED (X) https://github.com/rst-tu-dortmund/mpc_local_planner
-Teb Applied, TESTED (X) https://github.com/rst-tu-dortmund/teb_local_planner_tutorials
As you can check (Please) in Teb and Mpc there are launch files and examples to run it on Stage or Gazebo/rviz. So it is easy go through cfg and launch files to change and adapt to my simulation reality.
HOWEVER in the rsband and reed-shepp path planners algorithms, there are not examples, launch files or a complete README to understand how to apply these planners to Navigation Stack or how to declare the parameters in the move_base node launch file. Usually is just declare the global_planner “ROS_TYPE_Planner” and local_planner = “ROS_Type_local_planner” in params name and values in the node move_base.launch
Those github files contain a lot of navigation.h files in c++ that I do not know the process to embed them in navigation stack to work as local planners.
My C++ knowledge is still not good. Is there a video class or tutorial that explain how to use these kind of C++ planners algorithms described above or specifically the ones of OMPL (http://ompl.kavrakilab.org/planners.html) in Navigation Stack?
This question also was done by this guy and unfortunately ROS community did not answer:
It would be very very interesting if the Navigation Stack had the same interface of Moveit that allow to switch easily the algorithms and evaluate them through a benchmarking. If not possible directly to test these OMPL planners what is the other way , you experts suggest me please?
Thanks in advance