Robbyant has open-sourced LingBot-Touch, a tactile foundation model pretrained on nearly 5,000 hours of real tactile data that unifies magnetic, vision-based, and piezoresistive tactile signals. The model completed 26 real-world robotic tests on Sharpa Wave and XHand, and integrating it increased the average task success rate of four action models by approximately 14 to 26 percentage points. The technical report and code are available on Hugging Face and GitHub.
Source:roboticsandautomationnews.com · Read Original
Leave a Reply