Skip to content

feat(limx-tron1): Tier 1 obstacle-nav skill - #71

Open
Wanbogang wants to merge 1 commit into
fabricfoundation:limx-tron1-tier-1from
Wanbogang:feat/limx-tron1-tier1-robopay
Open

feat(limx-tron1): Tier 1 obstacle-nav skill#71
Wanbogang wants to merge 1 commit into
fabricfoundation:limx-tron1-tier-1from
Wanbogang:feat/limx-tron1-tier1-robopay

Conversation

@Wanbogang

Copy link
Copy Markdown

LimX TRON1 — Tier 1 (simulator-only)

Adds a complete RoboPay skill profile for the TRON1 wheeled-biped robot:
MuJoCo scene with real actuators (2 legs, hip-knee-wheel each, live-computed
rolling gait), fail-closed Zenoh bridge (robot/tunnel/action ->
robot/tunnel/result), result-gated settlement, and replay protection.

Sim-to-Sim validation

The same potential-field navigation policy runs in both MuJoCo and Webots
against the same obstacle/goal layout. PASS on all checks -- see
docs/evidence/sim_to_sim_validation.json.

Evidence

  • 11/11 contract tests passing (tests/test_bridge.py)
  • Success run: goal_reached, 0 real collisions, settlementEligible=true
    (docs/evidence/terminal_log.txt, tron1_metrics.json)
  • Deliberate failure run: timeout, settlementEligible=false
    (docs/evidence/failure_terminal_log.txt, tron1_failure_metrics.json)
  • Replay of the same actionId is rejected with no second simulator episode

See docs/README.md and docs/validation-report.md for full architecture
and reviewer verification checklist.

@Wanbogang
Wanbogang requested a review from a team August 5, 2026 05:38
MuJoCo scene for the wheeled-biped LimX TRON1: 2 legs (hip-knee-wheel each)
+ planar base, mass 14kg, live-computed rolling gait (wheel spin
proportional to commanded velocity, not a canned trajectory).

Bridge follows the same fail-closed pattern as my Deep Robotics
submissions: robot/tunnel/action -> robot/tunnel/result, result-gated
settlement, SQLite replay guard. 11 passing tests, plus success and
timeout demo runs with real evidence.

Includes Sim-to-Sim validation: the same navigation policy runs in both
MuJoCo and Webots against the same obstacle/goal layout. PASS on all
checks (see docs/evidence/sim_to_sim_validation.json).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant