Home

MyoWebToolkit is a tool for studying the electromyogram of the forearm with respect to the hand movements in order to reverse engineer the whole process, that is to estimate hand movements from the electromyogram. Most of the approaches treat this problem as a traditional pattern recognition problem, i.e. fitting a linear or non-linear classifier to signal features given the annotated output, which however we consider not a true solution due to the lack of a theoretical background to the certain problem and due to the derivation of discrete labels whereas the problem is of analogue nature. Therefore, we have attempted a solution that derives a) 15 muscles activation levels from the electromyogram of Myo device of 8 channels; b) conversion of 15 muscle activation levels to muscle forces; c) application of forces to a bones model in WebGL 3D governent from a physics engine. These 3 pillars constitute the backbone of the problem for real-time emulation of hand movements in an analogue form. The MyoWebToolkit is complete written in Javascript - PHP language to allow emulation in the client side (web browser) using the GPU. A complementary tool for recording myo data is also written in node.js technology that sends from every client data to the server in order to collect a database.

What is MyoWebToolkit?

MyoWebToolkit is targeting to researchers that want to experiment with Myo device and various algorithms for reverse engineering the hand movements. It can be used for demos, visualization, remote recording and other.

It differs from existing research solutions because it runs in a Chrome client (web browser) using javascript. It is coded remotely and it is always updated.

Concept

Muscle intersections

Future work

Recording experiments

Finger movements can not be easily detected by Myo. A database should be constructed to analyse 5 finger flexions and 5 fingure extensions, ten in sum, on various class cases.

Classes

1. No load movements: Perform the ten aforementioned finger movements in the air, plus one relax state.
2. Keys press movements: Perform 5 flexion finger movements (e.g. at computer keyboard, to press 1,2,3,4,5 and space keys), plus a relax state on the keyboard without pressing any key.

IMUs are not connected with Simulation tool

ToDos: connect the simulation tool with IMUs. Now the hand moves with W,A,S,D keys. In the current version 0.8 only EMGs are connected.