TacClaw-Infinity: A Large-Scale Vision-Tactile-Language-Action Dataset for Dexterous Manipulation

Key points

• Large-scale robot learning increasingly relies on diverse human demonstrations, yet tactile manipulation data remain difficult to scale, and existing multimodal datasets often lack consistent quality control and structured supervision.1
• TacClaw-Infinity is built through a distributed worldwide collection across homes, laboratories, industrial workplaces, retail stores, and outdoor environments including tasks contributed by academic and industrial collaborators.1
• We develop a scalable pipeline for trajectory reconstruction, multimodal quality control, embodiment-aware filtering, and structured annotation, covering contact-centric segmentation, subtask description, and spatial grounding.1
• We validate TacClaw-Infinity at both the data and model levels: At the data level, we assess the collection interface, the action-space compatibility of the curated demonstrations, and the structured annotations, while for the model level, we demonstrate cross-embodiment transfer across three heterogeneous robots without robot-specific demonstrations.1

Sources (1)

• [1]TacClaw-Infinity: A Large-Scale Vision-Tactile-Language-Action Dataset for Dexterous Manipulation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 03:06 PM

  We present TacClaw-Infinity, a large-scale structured Vision-Tactile-Language-Action (VTLA) dataset for contact-rich manipulation, collected with TacClaw, a tactile-enabled UMI-style handheld interface that records synchronized visual observations, fingertip tactile signals, and end-effector trajectories without requiring a target robot during data collection.

  Large-scale robot learning increasingly relies on diverse human demonstrations, yet tactile manipulation data remain difficult to scale, and existing multimodal datasets often lack consistent quality control and structured supervision.

Extractive summary: sentences quoted from the sources.