Tag
1X has developed new 25-DoF robotic hands for its NEO humanoid platform, achieving human-level dexterity, tactile sensing, and durability for real-world manipulation tasks.
OmniTacTune introduces a two-stage reinforcement learning pipeline for adapting tactile feedback to pretrained visual robot policies, achieving 85-100% success on contact-rich manipulation tasks within 40-80 minutes.
Presents Splash, a mask-isolated tactile alignment learning framework that enables multimodal LLMs to acquire tactile sensing without sacrificing vision-language reasoning by selectively updating only dormant parameters, preventing catastrophic forgetting.
Researchers introduced T-Rex, a framework that integrates vision, language, and tactile sensing, enabling robots to respond to physical contact in real time rather than relying solely on vision.
Researchers from CMU and Bosch Center for AI introduced the Humanoid Transformer with Touch Dreaming (HTD) model, which uses tactile signal prediction to improve humanoid robot manipulation, achieving a 90.9% higher average success rate over the ACT baseline across five real-world tasks.