Parkinson's is a debilitating disease that causes patients to gradually lose fine motor skills due to hand tremours. Our goal is to develop a cost effective and long-term solution to control the tremours - and more importantly, give back independence to patients.
What it does
Steady Hands uses electrical muscle stimulation to dynamically counteract Parkinson's tremors in the hands. There have been several proof-of-concept studies published in the past 10 years that demonstrate using electrical muscle stimulation to successfully diminish hand tremors in Parkinson's patients and thus improve their quality of life. However, ours is the first commercially viable implementation.
How we built it
We used a Myo Gesture Control Armband and used Lua and C++ to exploit the armband's haptic feedback feature - which cannot normally be controlled through API - in order to modify it into an electric muscle stimulation device. We also created an algorithm that distinguishes between normal hand movements and Parkinson's tremours from the multiple sensor feedback from the Myo armband. Hence, the Myo armband will sense tremours in a muscle, and then apply electrical stimulation to the complementary muscle to counteract the tremour and thus stabilize the hand. To develop our algorithm, we performed analytics with R from our own dataset collected from the Myo armband.
Challenges we ran into
There was no direct way of controlling the haptic response feature on the Myo armband through their API. We also had no way of differentiating between Parkinson's tremours and normal arm and hand movements.
Accomplishments that we're proud of
We were able to go under the hood and modify the way that Myo armband works to control the electrically produced haptic response in order to devise a makeshift electrical muscle stimulation device. We're also proud to have been able to perform statistical analysis on the sensor data to produce an algorithm that can distinguish between normal movements and tremours.
What we learned
We learned several new technologies, especially Lua and the Myo armband hardware. We also played around with MongoDB Stitch, although we ended up not using it for our final hack.
What's next for Steady Hands
We're looking forward to further optimizing our algorithm with machine learning and potentially trying it out on Parkinson's patients.