The entire system, which was developed in cooperation with ETH Zurich, has a modular structure and consists of three subsystems: a mobile robot, an electric robot arm and the 'BionicSoftHand 2.0'. This pneumatic gripper is inspired by the human hand - Festo presented an initial version in 2019.

To enable the 'BionicSoftHand 2.0' to perform the movements of the human hand in a natural way, small valve technology, sensors, electronics and mechanical components are integrated in a very small space. The fingers and opposable thumb consist of flexible bellows structures with air chambers, encased in a firm yet flexible knitted textile. This makes the hand light, adaptable and sensitive, but still capable of exerting strong forces. The pneumatic fingers are controlled via a compact valve terminal with piezo valves, which is attached directly to the hand.

In order to increase the range of movement of the thumb and index finger compared to the first version of the 'BionicSoftHand', the developers have significantly increased the lateral swivel range of both fingers. As a result, they can now work together optimally and grip very precisely. Thanks to a 3D-printed wrist with two degrees of freedom, the hand can also move forwards and backwards as well as to the left and right. This also enables gripping with a tight radius.

For more stability in the fingers, there are now two structural elements in each of the air chambers that act as bones. A bending sensor with two segments per finger determines the position of the fingertips. The hand also wears a glove with tactile force sensors on the fingertips, the palm and the outer sides of the robotic hand. This enables it to feel the properties of the gripped object and adapt its gripping force to the respective object - just like us humans.

In addition to tactile sensors, the hand has a depth camera on the inside of the wrist for visual object detection. With its help, the robotic hand can recognize and grasp various objects, even if they are partially covered. After appropriate training, the hand can also use the recorded data to assess the objects and distinguish good from bad, for example. The information is processed by a neural network that has been trained in advance with the help of data augmentation.

The 'BionicSoftHand 2.0' is combined with a mobile ballbot and a lightweight, electric robotic arm - the 'DynaArm'. It enables fast and dynamic movements. This is thanks to its lightweight design with highly integrated drive modules weighing just one kilogram.

The developers rely on an ingenious drive concept for the Ballbot: the robot balances on a ball. This allows the 'BionicMobileAssistant' to maneuver in any direction. The system has its entire energy supply on board: the battery for the arm and robot is located in the body; the compressed air cartridge for the pneumatic hand is installed in the upper arm. This means that the robot is not only mobile, it can also work autonomously. The algorithms stored on the host computer also control the autonomous movements of the system. With the help of two cameras, the robot orients itself independently in space.

The system would be predestined for use as a direct assistant to humans, for example as a service robot, as a helping hand in assembly or to support workers in ergonomically stressful or monotonous tasks. They could also be used in environments where humans cannot work, for example due to hazards or limited accessibility. Maintenance or repair work, the measurement of data or visual inspections are particularly conceivable. It is also conceivable that mobile robots could take on the simplest tasks in areas where there is an increased risk of infection or a lack of personnel due to infection. For example, a possible future scenario could be that they bring drinks and food to tables in restaurants or deliver medication to hospital patients or people in need of care in retirement homes.

Thanks to its modular concept, the 'BionicSoftHand 2.0' can also be quickly mounted on other robot arms and easily put into operation.