Pause for a moment and figure this out – you apply different levels of pressure when it comes to picking up different objects, for example, you will not grasp an egg in the same manner as with a rock. Robots, however, do not have the luxury of such experience, and they need to be programmed in the past to apply different levels of pressure when picking up different objects. Well, the HRP-3 Promet Mk.II might be a breath of fresh air, as it will rely on visual recognition to get the job done, such as the video which you can view after the jump, showing it hold a bottle with its multi-fingered hand based on visual recognition. One camera will recognize the object, where it will then determine the object’s size and position in space. No idea on how robust can it be, but hopefully it can tell the difference between something soft and hard.