While the use of prosthetic arms has become easier with the advancement of technology, thanks to artificial intelligence software, movements are now much faster and closer to reality. Prosthetic arms, which now move faster, also provide a sense of touch and movement in space thanks to the sensors on them.
Prosthetic arm and hand technology has developed a lot in recent years. Hands with fingers and joints that move independently of each other provide great convenience for prosthetic hand users. Such prostheses can be customized with 3D printers. Such prostheses utilize the activity in the arm muscles and use the signal from the muscles to operate the prosthesis. However, artificial hands still pose some challenges for users. Most users find it difficult to control prosthetic arms and hands.
The problem is that users often have to contract their muscles in certain combinations of patterns to create hand or wrist movements. With this combination, it takes time for the prosthesis to recognize and activate the movement. In addition, sometimes the prosthesis cannot perform the desired movement.
Underlying this problem is the fact that the signals to the brain for hand or wrist movements are still not fully classified. This makes it difficult for scientists to interpret the signals correctly and make prosthetic hands and arms that work as intended.
Researchers are now trying to solve the problem by adding artificial intelligence to the work. Diu Khue Luu and Anh Tuan Nguyen from the University of Minnesota, together with their colleagues, have taken an important step towards solving this problem by using artificial intelligence software that learns the user’s purpose from the nerve signals it detects from the arm. “We use artificial intelligence software to understand what movement the disabled person wants to do and offer a brain-powered system,” the scientists say.
ARTIFICIAL ARM LEARNS MOVEMENTS
To teach the AI the movements of the hand, the user needs to wear a special glove that transmits data to the good hand. The other hand is fitted with a signal collecting device. While the user performs all the movements with his/her good hand wearing the glove, he/she tries to perform the same movements with the other hand. The artificial intelligence matches and records the movements made by the intact hand with the data coming from the device attached to the other side. When all the movements are learned by the artificial intelligence, the prosthetic arm is attached and the device and artificial intelligence are paired via bluetooth.
When the prosthetic arm is to be moved, the artificial intelligence adapts the signals from the device on the amputated arm to the previously recorded movements. Thus, the prosthetic arm makes the desired movement much faster and more accurately. As the user performs the movements, the prosthetic arm grasps the movements more accurately. Luu, Nguyen and their team say, “Artificial intelligence decodes users’ movement intentions, allowing them to control their upper body prosthetic limbs with their thoughts.” Thanks to the sensors on the prosthetic arm, it also provides the user with a sense of touch and movement in space.
One of the companies developing prosthetic arms is New York-based Esper Bionics. The prosthetic arm and hand called Esper Hand produced by the company also works with artificial intelligence support. Esper Hand can receive signals from individual muscles and control objects three times faster than prostheses without artificial intelligence support. The 380-gram prosthetic arm is made of composite material. Controls are done through a wearable brain-computer interface. A cloud-based artificial intelligence software personalizes the control of the hand and ensures that the prosthesis works properly.
“As we develop the Esper product, we always remember that this is about rewriting millions of life stories, helping people live more fulfilling lives,” says Dima Gazda, CEO and co-founder of Esper Bionics.
In prosthetic arms powered by artificial intelligence, the more the user uses the arm, the better the prosthetic arm performs the movements. The prosthetic arm can detect the behavioral patterns that the user repeats the most and thus predict the user’s movements. This helps train the arm to move intuitively without any effort from the user. It can also be controlled and trained through a mobile app.
Sources: Time, ZDnet, Discovermagazine