Where is MLI™ today?
From Søren’s innovative idea of a system capable of monitoring, evaluating and informing the users to help them in their behavior modification, we began the product development 3 years ago. We knew the cause of the issue and that it is mainly related to muscle activity. We sought high and low to find a way to properly monitor the muscular activity and it soon became clear that we needed to measure the muscle’s bio-electric signals, also known as electro myography, EMG. Which is method to sense and measure the amount of activity that is present in the muscle to perform what we see as a lifted finger, us grapping something or a moving wrist.
With our first funding we developed our first mock-up – the very first concept of our design.
But was it any good?
We made a study where we compared the obtained signals and we found that we had succeeded in developing a device that could collect EMG. We were happy but soon we were challenged!
How good was it at measuring the EMG then?
Luckily, we got additional funding which allowed for the development of a new set of prototypes. We performed a “Gold Standard comparison test” against a medical device, real life, monster system of an EMG measurement system at a hospital. With the help from the hospital and a university we got the test performed and data crunched. The conclusion was clear – even though, our prototype was a smaller system, with less resolution of the measurement, our prototype could run up with the laboratory EMG measurement equipment.
The first step had been established! We could measure the muscle activity! We could monitor!
But what about the evaluation? Could we say anything about the muscular activity, hence, the user’s behavior?
With the help from additional funding we had more prototypes build and did a controlled pattern test. The test was performed with 80 people who did the exact same 26 tasks; lifting fingers, lifting weights, bending the wrist, scrolling a mouse, etc. Again, with help from the university the data was crunched and it very soon became clear, that not only could we see differences in loads, e.g. lifting a 2kg weight or a 5kg weight, we could from the signal identify, with very good confidence, which of the 26 tasks each and every user was performing, simply by looking at relevant parameters from the signal.
This was a great result because it meant that the quality of the data was sufficient to do advanced evaluation upon. Additionally, it also meant that with such good data we would be able to evolve the evaluation with the help of machine learning. We would be able to make better and better feedback based on the data we would collect.
But how to evolve the algorithms? How to make our feedback better?
We needed data, we needed test persons, and we needed a lot of them. Again, we got more funding!!! We had more prototypes made and we released our second version of our app in Google Play and Apple App Store. Throughout 3 months we did nothing but perform tests at volunteering companies. 34 companies volunteered, with more than 450 employees involved in the data collection, where the employees wore our product for 2 days. This resulted in more than 6000hours of quality data being collected. Everything was stressed! The hardware, the app, the cloud solution, the data processing. But we succeeded and we learned so much! The feedback to the user began to make shape; the Muscle Load Index™. Our own, PRECURE-made aggregated index, that provides the knowledge of how strained the muscle is – vital information for the user on the journey of reducing their strain, thus, reducing the risk of getting a strain injury.
The inputs from this test is right now being built into the hardware, the app and the cloud solution, fine tuning the product, and we are very close to our approved final product, where we will launch November 2019 at our website.