The entire system has been developed in collaboration with ETH Zurich. It has a modular design consisting of three subsystems: a mobile robot, an electric robotic arm and a pneumatic gripper – the BionicSoftHand 2.0 – an earlier version of which was first presented by Festo in 2019. It uses a combination of compact valve technology, sensors, electronics and mechanical components to emulate the movements of a human hand.
BionicSoftHand 2.0 wears a tactile glove with force sensors on the fingertips, the palm and the outer sides of the robot hand. This allows it to sense the nature of the item to be gripped, and to adapt its gripping force accordingly. With a 3D-printed wrist with two degrees of freedom, the hand can also move both back and forth, as well as to the left and to the right.
In addition, the hand has a depth camera on the inside of the wrist so it can visually detect the objects to be gripped – even if these are partially obscured. The information is processed by a neural network, which has been trained in advance with the aid of data augmentation. Once the hand has been correctly trained, it uses the intelligence gathered to assess objects.
Festo says that the system would be perfect for use as a personal assistant, for example, as a service robot helping in assembly or assisting workers in carrying out ergonomically monotonous tasks, or for differentiating good product from bad product. It could also be used in environments where people cannot work because of hazards or restricted accessibility. This includes maintenance or repair work, data measurements or visual inspections.
To see the BionicMobileAssistant in action, visit the Festo website.