search

NASA-Tensegrity-Robotics-Toolkit / NTRTsim

The NASA Tensegrity Robotics Toolkit Simulator, a physics based simulator to research the design and control of tensegrity robots.
Apache License 2.0
161 stars 81 forks source link

Contact Cables Cannot Wrap #119

Open brtietz opened 9 years ago

brtietz commented 9 years ago

If a tgBulletContactSpringCable gets wrapped a full turn around an object, it immediately loses all of its anchors. See muscleNPTests demo in my (btietz) dev folder. This is likely due to the anchor sorting scheme in the anchorCompare struct. While not a major problem for tensegrity structures, a more sophisticated sorting scheme would allow for wrapped cables to simulate motor spools and other aspects.

brtietz commented 9 years ago

Upon further investigation, the cables can wrap if one does not delete based on the results of the anchor contact testing (see lines 518 and 564 of tgBulletContactSpringCable.cpp as of commit 83f492c4158d6b562ef9529c56b3f38d9bb422a4). However, then anchors never lose contact with their respective objects based on sliding off. If we can determine a scheme for lost contact that does not involve btPersistentManifolds, we should be able to make progress on wrapped cables.

vsunspiral commented 9 years ago

this might be where dijkstra's algorithm helps out. basically see if the straight line path between the prior anchor and next node after the last attached anchor are in free space. we would need a 3d version of dijkstras (it is normally 2d) which I'm sure exists (and has some other name).

http://en.wikipedia.org/wiki/Dijkstra%27s_algorithm

vytas

On Jan 7, 2015, at 6:58 AM, Brian Mirletz notifications@github.com wrote:

Upon further investigation, the cables can wrap if one does not delete based on the results of the anchor contact testing (see lines 518 and 564 of tgBulletContactSpringCable.cpp as of commit 83f492c). However, then anchors never lose contact with their respective objects based on sliding off. If we can determine a scheme for lost contact that does not involve btPersistentManifolds, we should be able to make progress on wrapped cables.

— Reply to this email directly or view it on GitHub.

Vytas SunSpiral
Dynamic Tensegrity Robotics Lab cell- 510-847-4600 Office: 650-604-4363 N269 Rm. 100

Stinger Ghaffarian Technologies Intelligent Robotics Group NASA Ames Research Center

I will not tiptoe cautiously through life only to arrive safely at death.