Closed ayoubraji closed 5 years ago
The RRTXPlanner is a prototype for a global planner as plugin for nav_core and the rrtxnode as a standalone node for ROS.
I've tested this code and the ompl rrtstar on a model car. The rrtstar code is more optimized (obstacle detection on edges only if they are more optimal) and therefore faster in practice.
http://wiki.ros.org/nav_core http://ompl.kavrakilab.org/classompl_1_1geometric_1_1RRTstar.html https://github.com/AutoModelCar/AutoModelCarWiki/wiki/Hardware-(AutoNOMOS-Model-v3)
Now it's clear. The rrtstar is optimized but I'll try the rrtx because I'm searching a solution for a dynamic environment, in which obstacles could be in movement. Thank you for the references.
Thank you for the openness
I'm checking the code because I'm interested in rrtx for a dubins car like robot planning and obstacles avoidance. I don't understand the purpose of the RRTXPlanner.hpp and cpp code. Are they only part of an initial attempt for the rrtxnode? Because I've seen that the rrtx_node works only thanks to the main rrtx algorithm and the publisher.
Furthermore, have you tested the code in simulation or in a car like robot? Thank you