Students and scientists from Vladivostok have developed a prosthesis that learns and adapts to the patient using a neural network.
The students made a lower limb prosthesis. In standard such products, the repulsion of the leg from the surface is compensated by a special spring structure. They cannot compensate for the angle of elevation of the foot, which makes the prosthesis uncomfortable.
The new prosthesis has the ability to swing the foot and compensation for repulsion. Thanks to this, a person saves resources when walking and “thinks a little about the positioning of the foot”. The prosthesis must also receive external signals in order to “know” at what point the foot should push off the surface. This is the question that neural networks have solved.
The scientists used a sensor system that was attached to the leg of a healthy person. So they got the data for training the neural network. As a result, the algorithm predicts the angle of rotation of the foot. The prosthesis can also be adapted to the individual characteristics of the gait.
Photo by RIA Novosti was used for decoration
Read also:
Doesn't lag or freeze: remember this inexpensive smartphone with a super-powerful processor
The outcome of the duel between the Russian Su-24M bomber and the British destroyer is revealed
Russia will start selling the “killer” of the Turkish drone Bayraktar abroad
Scientist named the main goal of the coronavirus
The latest submarines will be “easy prey” in the war
