Robot Data and Hardware

Prevent robot evaluation leakage with episode, session, site, object and operator group splits, sibling detection, train-only preprocessing and frozen tests.

Mine robot failures from rollout windows using uncertainty, novelty, impact, clustering and human review, then recollect safely and retest holdouts.

Collect human-in-the-loop robot intervention data with control authority, pre-takeover context, correction intent, timing, safe handback and evaluation.

Weight robot dataset mixtures by domain and outcome while controlling sampling exposure, oversampling, episode length, drift and worst-domain performance.

Normalize robot actions across embodiments while preserving command type, frames, units, gripper meaning, timing, masks, statistics and round-trip tests.

Read ISO 18646-5:2026 legged robot performance results by checking edition, test machine, conditions, metric definitions, variation and safety scope.

Understand LeRobot Dataset v3 Parquet, MP4 and metadata shards, then validate episode offsets, timing, finalize, streaming and v2.1 migration.

Adapt humanoid walking to stairs and slopes with geometry estimation, full-sole support, swing clearance, landing verification, replanning and tests.

Design humanoid fall detection and protective motion with recoverability, direction-specific contact, impact metrics, staged tests and safe restart gates.

Detect humanoid foot slip from force and motion, protect state estimation, then recover by unloading, momentum control, stepping or safe stopping.