Can prosthetics be intelligent? Yes, they can. “Intelligent prosthetics are prosthetics that perceive their surroundings via sensors. Based on these perceptions, they then adapt their functions appropriately to meet the patients needs,” explains Professor Arndt Schilling, prior to his presentation at MEDICA 2018.

Schilling is the head of research and development at the clinic for trauma surgery, orthopaedics and plastic surgery at the University Medical Center Göttingen. For him, a typical example of intelligent prosthetics are the “mind-controlled” prosthetics. They pick up nerve signals, derive actions from them such as “open prosthetic hand” and carry them out. The biggest challenge here is to reliably interpret the patient’s intentions. Schilling explains, “Thoughts will always be free and are therefore extremely hard to trap in algorithms. And it is also sometimes difficult for developers to understand the patient’s complex, changed everyday situation, especially because the developers generally do not have prosthetics themselves.” To be able to adapt intelligent controls to the patient’s needs, extremely close cooperation in an inter-professional team of patients, doctors and engineers is therefore necessary.

Prosthetics learn how to live with the patient, not the other way around

Prof. Arndt Schilling, head of research and development at the clinic for trauma surgery, orthopaedics and plastic surgery at the University Medical Center Göttingen. (c) University Medical Center Göttingen.

Machine learning can help in developing such prosthetics and orthotics. Until now, conventional controls meant that patients had to train for a long time before they learned how to carry out movements in a way the prosthetics could understand, says Schilling. He explains, “Machine learning enables patients to carry out movements in ways that make the most sense to them, and the prosthetics train themselves to understand the patient.” Machine learning reverses the teacher-pupil-relationship between patient and prosthetic, so to speak. “This gives patients the pleasant feeling of the prosthetics serving them and not the other way round,” says Schilling.

Combining machine learning with extended sensors could also give prosthetics a basic understanding of their surroundings. Schilling explains, “In our InoPro project we are working on prosthetics that can recognise whether the patient is reaching for a glass or a pencil, for example, and prepares the hand’s position and grip accordingly, which in turn relieves the patient.”

Today, first prosthetics have been equipped with relatively limited intelligence that allows them to adapt mechanics to the respective gait phase or regulate the speed at which an object is reached for. “The first start-ups are looking into this subject and I expect similarly quick progress to that of voice control in the past years,” says Schilling.

The fundamental technical difficulty is due to the fact that the natural human motor system is so amazingly well-engineered. “Trying to develop something that is even remotely like a hand or a foot is automatically high tech.” Intelligent prosthetics not only have to be smart, they also have to be robust, light and waterproof. Ideally, they should also require very little power, so that they do not have to be recharged all the time. The sensors need to work reliably, no matter whether the wearer is moving, freezing cold, or sweating. This requires biocompatibility of the used materials and the surface of the body.

Ethical border case: Does everyone have a right to high tech?

Currently, Schilling primarily sees an ethical border in the discussion on who should be given access to these modern aids, and who shouldn’t, as high tech is of course very expensive. “Does everyone have the right to be provided with state-of-the-art prosthetics, should the need arise? Which standard will supportive society accept? How much do we want to invest in further development? How can we ensure that people outside of our supportive society benefit from these developments?” Schilling believes the answers to these ethical societal questions will fundamentally determine future developments in this field.

Don’t miss:
Wednesday November 14, 11:00-11:20 a.m.
Session 7: Visioneering the future of healthcare: Robotics in healthcare – Neurostimulation – VR / AR in healthcare
Keynote: Connecting robots, sensors and machine learning – The future of bionic technologies
Prof. Dr. Arndt Schilling, Head of Research and Development, University Medical Center Göttingen

Source: MEDICA