| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
Can someone shed some light on why A* star is used for global planner in move_base. I can see that there is an option to use Djikstra's algorithm as well. What are the main benefits of A* as opposed to using RRT or RRT*?
Outside of RRT, here are my thoughts on using A* vs Djikstras
This is a good basic reference for A* vs Dijkstra. A* has an added heuristic trying to expand the potential field in the direction of the goal. Its not necessarily promised to be faster, but in most cases it is.
http://theory.stanford.edu/~amitp/Gam...
RRT and RRT-variants are sampling based and not search based. They will require smoothing after the search. In practice RRTs are nice for higher dimensional spaces like manipulators, but in lower dimensional space (like 2D mobile robots) its not really adventitious over search based. WIth that said, there are RRT implemented global planner plugins if you google for it for you to try out youself if you like.
Asked: 2019-09-11 10:52:09 -0500
Seen: 367 times
Last updated: Sep 11 '19
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
There is a Stackoverflow post about A* here take a look at that, mybe this gives a bit of an insight, also here you can find a comparison between Dijkstra and RRT.