plannerBiRRT seems to extend search tree only in one direction?!

6 Ansichten (letzte 30 Tage)
Hello everybody,
i need to plan a path for a robot manipulator. The movement should be done between sections at a constant operating height. (see Picture) I used this example to realize it: https://de.mathworks.com/help/robotics/ug/plan-paths-with-end-effector-constraints.html
The start is the big red dot on the far left and the goal is the big green one. As you can see does the plannerBiRRT take the longer way around the left corner but it should in my opinion find the shorter path around the right corner faster and be able to move that way. The positions on this way are definitly reachable for the robot. But it seems that the start and goal tree both extend in the left direction. I've also looked into the the plannerBiRRT class and it's superclass matlabshared.planning.internal.BiRRT. I've found a variable extendsReversly in line 262 of the superclass code. Could this be the reason why? Or are there any other reasons why the trees seems to move/search only in one direction.
I would appreciate any information that solves or explains this problem!
Thank you very much!

Akzeptierte Antwort

Fabian Vacha
Fabian Vacha am 23 Jan. 2025
Bearbeitet: Fabian Vacha am 23 Jan. 2025
For anyone wondering, i solved this problem a few months ago.
It was because the bounds of the y- and x-rotation of the state space were constrained to zero. After i gave the state space full rotational freedom from -pi to pi for the z-, y- and x-rotation it worked.

Weitere Antworten (1)

Kaustab Pal
Kaustab Pal am 11 Jul. 2024
The performance of the plannerBiRRT can be influenced by various parameters. You can try increasing the value of the "MaxConnectionDistance" parameter, this could resolve the issue you are encountering. I confirmed this through testing the provided example within the plannerBiRRT package, which is accessible here: Create bidirectional RRT planner for geometric planning - MATLAB - MathWorks
By default, the generated path tends to be longer. However, upon adjusting the "MaxConnectionDistance" value to 0.3, a shorter path was successfully obtained.
Path generated with default parameters:
Path generated after increasing “MaxConnectionDistance” from 0.1 to 0.3:
Hope this helps.
  1 Kommentar
Fabian Vacha
Fabian Vacha am 15 Jul. 2024
thank you for your answer! Unfortunately this doesn't solve the problem. The planner still takes the longer way :(

Melden Sie sich an, um zu kommentieren.

Kategorien

Mehr zu Robotics finden Sie in Help Center und File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by