Conversation
TomCC7
force-pushed
the
feat/behavior-integration
branch
from
September 16, 2026 17:55
4dccd8c to
542d0d2
Compare
TomCC7
added this pull request to stack #4160
September 16, 2026 17:56
❌ 13 Tests Failed:
View the top 3 failed test(s) by shortest run time
To view more test analytics, go to the Test Analytics Dashboard |
TomCC7
force-pushed
the
feat/behavior-integration
branch
from
September 17, 2026 23:36
8968a43 to
4244ab1
Compare
TomCC7
force-pushed
the
feat/behavior-integration
branch
from
September 18, 2026 00:35
4244ab1 to
b36b1e7
Compare
TomCC7
force-pushed
the
feat/behavior-integration
branch
from
September 18, 2026 00:39
b36b1e7 to
2abc3c7
Compare
TomCC7
force-pushed
the
feat/behavior-integration
branch
from
September 18, 2026 01:50
2abc3c7 to
be5975f
Compare
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem and solution
Run BEHAVIOR/OmniGibson in DimOS without mixing Isaac Sim's Python 3.11 / NumPy 1 dependencies into the host environment. Connect the benchmark's 2026 R1 Pro to the existing navigation and manipulation stacks.
behavior-r1problueprint connecting KronkNav, ControlCoordinator and ManipulationModule with the native Isaac viewport and Viser controls.--no-deps. Update ABC/LeRobot callers and package runtime sources without local assets or environments.Integration fixes
The live proof required fixed-link robot self-filtering with cached Open3D distance queries, removal of MLS's 64-candidate start-snap cap, serialized manipulation shutdown, and Zenoh 1.10.1's loopback-discovery fix. The Nix pre-commit invocation uses Git filtering to exclude ignored simulator assets.
Run
After the one-time setup in
docs/capabilities/simulation/behavior.md:The combined blueprint opens the native Isaac viewport, native DimOS Rerun viewer, and Viser manipulation controls. Rerun shows the voxel map, traversable surface, robot heading, goal and path beside all three cameras. Click a mapped floor point to navigate; engage the Keyboard Teleop overlay for WASD/QE control and Space/STOP. Manual takeover cancels navigation and clears the path. The Linux launch command unsets WAYLAND_DISPLAY to avoid the observed viewer hang. The automated proof navigates, stops, executes planned motion with each arm, then cancels and checks position hold.
Validation
RUSTC=/usr/bin/rustc,RUSTDOC=/usr/bin/rustdoc) and Nix flakes enabled.Rs_int: all three RGB-D cameras, joints/TF, both arms/grippers, 0.090 m base-command motion, 0.157 m native-action motion, stable height and expired-command hold.Rerun control validation
Scope and remaining work
Localization uses simulator ground truth. Manipulation checks robot self-collision; the navigation map is not yet connected to its collision world. These proofs do not establish clutter-safe pick-and-place or official benchmark performance.
Earlier symbolic placement and physical grasp attempts failed; complete task success across reset and agentic execution remain unverified. The new R1 Pro navigation proof supersedes the earlier navigation failure. Keep this PR in draft.
Stacked on #4156 (
cc/feat/webxr-audio), above ABC #4088.Checklist
Agent: Codex (GPT-6)