0
$\begingroup$

I was reading the classical robotics paper Dynamically-stable Motion Planning for Humanoid Robots and I don't understand why the two trees are swapped in line 7 of figure 5 (page 11). Seems like an otherwise very straightforward algorithm, just this part is tripping me up.

Any help will be much appreciated, thanks!

$\endgroup$

1 Answer 1

0
$\begingroup$

I didn't read the paper fully, so I don't know if this was stated already, but the mechanism of swapping points would grow the path from both ends, which I think would lead to more graceful obstacle avoidance than just growing the path from A to B. Make a segment from A in the direction of B, then make a segment from B in the direction of A. Keep alternating this way and hope to meet in the middle.(again just my best guess)

Try contacting the authors! I did it once and was thrilled to get any response, and it was even helpful!

$\endgroup$
1
  • 1
    $\begingroup$ Chuck is exactly right. Section 3, page 5, says they "connect two search trees, one from the initial configuration and the other from the goal." That swap is how they grow two trees with the same algorithm. Reasonable software engineers might debate whether to implement it that way, but the algorithmic description is clear. $\endgroup$
    – r-bryan
    Nov 20, 2021 at 1:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.