Ant Group has open-sourced its tactile foundation model, LingBot-Touch, which was trained on nearly 5,000 hours of real-world tactile data. This model aims to enhance robots' physical perception and closed-loop manipulation capabilities by unifying tactile representations across different sensors and hardware. In tests, integrating LingBot-Touch improved task success rates by 14-26 percentage points for various robotic actions, demonstrating its potential to accelerate the development of embodied AI. AI
IMPACT Enhances robot dexterity and physical interaction capabilities, potentially accelerating the development of embodied AI applications.
RANK_REASON Release of a new foundation model for tactile sensing in robotics, with technical details and performance metrics. [lever_c_demoted from research: ic=1 ai=1.0]
- Ant Group
- GPT -6 Astra
- LingBot-Touch
- LingBot-Touch-Small
- LingBot-Touch-Tiny
- LingBot-VLA
- Sharpa Wave
- XHAND1
- Zhu Xing
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →