Planar control tuning, observation loss and transfer

01 / QUESTION

153 episodes and 306000 physics steps. The selected controller passes 7/18 holdouts; zero force coasts, and mass scaling lowers mean error without passing the velocity contract.

METHOD

Freeze the protocol, execute it, independently audit raw records, then recheck an export from the Git commit.

RESULT

153 episodes and 306000 physics steps. The selected controller passes 7/18 holdouts; zero force coasts, and mass scaling lowers mean error without passing the velocity contract.

OBSERVATIONS

153 episodes and 306000 physics steps. The selected controller passes 7/18 holdouts; zero force coasts, and mass scaling lowers mean error without passing the velocity contract.

CONCLUSION

Read the three question-led tutorials, including negative results and validation boundaries.

LIMITATIONS

Planar force-controlled simulation without gravity or contact; no arm, vision or hardware. Full VL01 and M4/M5 remain incomplete.

REPRODUCE

Each article provides a pinned revision, fresh output directory, run and audit commands. Execute the shared protocol once.

1. Engineering cases

   MuJoCo controller tuning: best candidate is not a passing controller

   Select on 54 development episodes, then check position error and stable completion on 36 held-out episodes.

   2026-10-1110 MINPublished
2. Engineering cases

   MuJoCo observation loss: zero force is not a stop

   Compare stale feedback, zero force and fading in 27 episodes to distinguish timeout from physical braking.

   2026-10-1110 MINPublished
3. Engineering cases

   MuJoCo mass transfer: can the same gains still work?

   Compare fixed and mass-scaled gains in 36 episodes, including lower mean error without a velocity-contract pass.

   2026-10-1110 MINPublished
4. PROJECT / 003

   Hohoo Embodied Experiments

   Two MuJoCo teaching scenes—pick-and-place and planar force control—record contracts, tuning, clearance and delayed-feedback outcomes, including failures.

   2026-09-28Experiment