Keep a 3mm collision buffer on free traverses - #296
Merged
Conversation
Unconstrained moves now plan with CollisionBufferMM set to 3mm, so free traverses hold their distance from the machine instead of grazing it at armplanning's 1e-8 default. Steps carrying a LinearConstraint keep that default — they have to close the last millimetres onto hardware. The no-spill level carry builds its own PlanRequest, so it gets the buffer explicitly. Without that, a no-spill step's clearance depended on the no_spill_carry flag: off, it fell through to planToRawPose and got the buffer; on, it routed to carryHeldLevel and got none. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Nicolas Palpacuer (NickPPC)
requested a review
from Vijay Vuyyuru (vijayvuyyuru)
August 27, 2026 17:26
Vijay Vuyyuru (vijayvuyyuru)
approved these changes
Aug 27, 2026
Nicolas Palpacuer (NickPPC)
added a commit
that referenced
this pull request
Aug 28, 2026
Tunes the value #296 introduced. The clearance still applies to every unconstrained move, including the no-spill level carry; steps carrying a LinearConstraint are unaffected and keep armplanning's default. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Nicolas Palpacuer (NickPPC)
added a commit
that referenced
this pull request
Aug 28, 2026
* Carry the milk bottle without the level-carry constraint The bottle is tall and grasped high, so the no-spill pose cloud may simply be unplannable for it — and an unplannable carry fails the order rather than degrading. Free-plan both of its carries (to the pour position and back to the fridge) until the poses are calibrated and we know what the arm can actually hold. The pour pivots are untouched: tilt speed stays on pour_vel_degs_per_sec and pour_acc_degs_per_sec2. Cup and glass carries keep no_spill_carry. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Lower the free-traverse collision buffer to 2mm Tunes the value #296 introduced. The clearance still applies to every unconstrained move, including the no-spill level carry; steps carrying a LinearConstraint are unaffected and keep armplanning's default. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Free traverses planned with armplanning's default
CollisionBufferMMof1e-8, letting the planner route the arm within a hair of the machine. Unconstrained moves now plan with a 3mm buffer; steps carrying aLinearConstraintkeep the default, since they have to close the last millimetres onto hardware.Also closes a gap in the no-spill carry, which builds its own
PlanRequest: a no-spill step's clearance used to depend onno_spill_carry— off, the step fell through toplanToRawPoseand got the buffer; on, it routed tocarryHeldLeveland got none. Both paths now match.Pivot and circular moves lack a
LinearConstrainttoo, but keep the tight default deliberately — they are the grinding, cleaning, and shake contact phases.Testing
make lint(0 issues) andmake testpass. No new tests — the change is planner configuration.Claude Code prompts used
If not can you adjust to apply it then too?"
🤖 Generated with Claude Code