New DECO model advances bimanual robot manipulation with tactile sensing

Researchers have introduced DECO, a novel decoupled multimodal diffusion transformer designed for bimanual dexterous manipulation. This system effectively integrates vision, proprioception, and tactile signals through specialized conditioning pathways. Alongside the DECO model, the team has released the DECO-50 dataset, comprising 50 hours of data for bimanual manipulation tasks collected on real dual-arm robots. DECO demonstrated superior performance, achieving a 72.25% average success rate in real-world evaluations, with a further 10.25% improvement when utilizing a lightweight tactile adapter. AI

IMPACT Enhances robotic dexterity and control by integrating multimodal sensory data, potentially improving performance in complex manipulation tasks.

RANK_REASON Research paper detailing a new model and dataset for robotics. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• DECO
• DECO-50
• Gotit.pub
• Hugging Face
• ScienceCast
• Xukun Liu

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →