Scientists from Russia’s Tomsk Polytechnic University are said to be developing a robotic arm prototype and its control algorithm using myoelectric signals. The mechanical limb would independently recognize the motions of its owner and perform all the same motions of a healthy arm, they reported.

According to developers Mikhail Grigoriev, Nikita Turushev and Evgeniy Tarakanets from the Institute of Non-Destructive Testing, the human prosthetic limbs have been available for a few decades, but to make them functional and translate them into a full replacement of a lost body part has been impossible.

“To date, there are quite (a number of) available traction prostheses,” Turushev said. “Their motions are carried out by means of traction belts which are superimposed from the repaired arm across the back (and) loop around (the) healthy shoulder. That is, the prosthesis performs by (using) motions of a healthy arm.” He added the drawbacks of that type of prostheses include that they are in need of unnatural body motions to control them.

The algorithm being developed by the polytechnic team will save people from having to wear traction belts, according to the researchers. Sensors on the prosthesis will pick up myoelectric signals. The human brain sends signals to muscles making them perform the necessary actions. The system will analyze commands coming to the healthy arm part and “guess” what motion the prosthesis should do.  

“Initially, software will be universal, but we will adapt it to each specific artificial arm,” Grigoriev said. “Further, a machine learning algorithm will copy its host wearing the prosthesis to fix myoelectric signals and choose required motions.”

Now the scientists are said to be “teaching” the algorithm different signals and along with the meaning of those signals. Initially, they plan to examine at least 150 people with healthy limbs.

The researchers received a grant from the Russian Foundation for Basic Research in 2015. In two years they hope to present the prosthesis prototype and its related software for operational support funding.

The scientists estimated the final cost of the device to be $600-1,000, although no details were provided as to how that expected price range was determined.