On November 10, Ant Lingbo Technology opened up the basic tactile model LingBot-touch to supplement the robot brain with key tactile perception capabilities. The model is based on nearly 5,000 hours of real tactile data pre-training, covering the three mainstream sensing principles of magnetotactile, visual touch, and piezoresistive. Different signals can be converted into the same set of tactile characteristics for use by different dexterous hands and motion models, so that the robot can “see clearly and understand” while further complementing the tactile perception required for fine operation.

On November 10, Ant Lingbo Technology opened up the basic tactile model LingBot-touch to supplement the robot brain with key tactile perception capabilities. The model is based on nearly 5,000 hours of real tactile data pre-training, covering the three mainstream sensing principles of magnetotactile, visual touch, and piezoresistive. Different signals can be converted into the same set of tactile characteristics for use by different dexterous hands and motion models, so that the robot can “see clearly and understand” while further complementing the tactile perception required for fine operation.