E3 / Contact monitoring during transfer

01 / QUESTION

Separate missed loss, false cancellation and confirmation delay using forced opening and one empty contact report.

METHOD

Fixed zero-offset scene; three conditions × three policies × three deterministic repeats = 27 rollouts. Observe at 50 Hz and hold the target for 0.6 s after an alarm.

RESULT

Immediate cancellation falsely interrupts one empty report; three-sample confirmation tolerates it. Both monitors cancel transfer after forced opening without recovering placement.

OBSERVATIONS

First bad sample: 4.802 s; three-sample alarm: 4.842 s. Confirmation adds 40 ms relative to immediate cancellation; subsequent measured XY paths are 11.902 and 22.534 mm respectively.

CONCLUSION

Monitoring changes actions after failure but does not recover the fallen cube; the threshold trades false cancellations against waiting.

LIMITATIONS

Privileged state, fixed conditions, transfer only. Forced opening is not natural slip. Path horizons differ; no vision, stale-packet timeout, recovery or hardware emergency stop.

REPRODUCE

11 boundary tests and 27 independent rollout audits passed; 18 matching state prefixes. Pinned CSV, full states, plots and two CPU coordinate replays. OpenGL scene rendering was unavailable.

1. Engineering cases

   Embodied practice III: a confirmed grasp still needs monitoring

   A 27-rollout MuJoCo comparison of transient observation loss and forced opening, with false cancellations, a 40 ms confirmation delay, code and recorded trajectories.

   2026-09-2914 MINPublished
2. Engineering cases

   Embodied practice II: if the grasp failed, skip the transfer

   An 18-rollout MuJoCo comparison of a lift/contact gate, with recorded trajectories, videos, runnable code and explicit limits.

   2026-09-2912 MINPublished
3. PROJECT / 003

   Hohoo Embodied Experiments

   Seven MuJoCo teaching experiments, from pickup bias to stage contracts, with raw trajectories, decisions and independent audits.

   2026-09-28Experiment
4. LAB / 006

   E2 / Confirm the grasp before transfer

   Use sustained height and bilateral contact to decide whether transfer should proceed.

   2026-09-29Completed