The Zhitong Finance App learned that on November 10, Ant Lingbo Technology opened up the basic tactile model LingBot-touch to supplement the robot brain with key tactile perception capabilities. Based on nearly 5,000 hours of real tactile data pre-training, LingBot-Touch covers the three mainstream sensing principles of magnetotactile, visual touch, and piezoresistive, and can transform different signals into the same set of tactile characteristics, so that the accumulated “feel” can be reused across sensors, dexterous hands, models, and tasks. In 26 real machine tests, after incorporating this characterization, the average task success rate of the four motion models increased by about 14 to 26 percentage points compared to when they were not tactile.

Previously, Ant Lingbo's LingBot series models had built a full-stack technology system covering high-precision spatial perception, high-dynamic environmental interaction, and generalized dexterous operation, so that robots can see more clearly, think more clearly, and work better. This time, LingBot-touch extends the sensing ability of the general brain to the “fingertips”, making the robot “more sensitive to touch” and further improving the ability of the native general brain.

Disclaimer:Webull uses external vendor Google Translation Service for news translations where we endeavour to ensure these are correct, however, we recommend that you please double-check this information accordingly. Webull is not responsible for translation errors or issues.