Prosthesis powered by Google’s TensorFlow AI allows amputee drummer to play again

When drummer Jason Barnes had the lower part of his right arm amputated following an accident at work in 2012, he was certain that it would put an end to him being able to play the drums. But now, Google has released a video (above) demonstrating the power of its machine learning technology, powering a prosthesis which allows Jason to play again.

Jason has been working with Gil Weinberg, a roboticist and founder of Georgia’s Tech Centre for Music Technology since 2013 to develop the prosthesis, which is operated via electromyography (EMG). 

The movements and gestures from the residual part of Jason’s limb generates electrical signals, which is converted into streams of data for the TensorFlow software to learn and process. 

“I can flex my muscle and it will tighten the grip on the stick, and I can extend the muscle to loosen the grip on the stick, just like it would on a normal hand. I can actually feel the feedback from the arm, and it feels as close to a real hand as you can get without it actually being a real hand.”

The result is that the prosthesis is able to react to Jason’s desired movements with minimal latency, producing some extremely natural control over the striking.

“I felt lucky to be a part of this process. In the past, there has always been a learning curve with me having to adjust to a new device. But this way, it was the device that was adjusted to me.”

You can follow Jason’s journey, as well as check out his musical projects via his Instagram account.