ARlimb is a free and open source immersive augmented reality system for the training of pattern classification control with a myoelectric prosthesis.

Our article describing the system was submitted for publication at the Journal of NeuroEngineering and Rehabilitation (JNER) in September 2020. To enable further research, the source code and the design of our system along with extensive documentation and ready-to-use examples are published under a free and liberal license on the project's website https://github.com/arlimb/arlimb. The modular design enables easy replacement of single components or integration in other projects. The hardware components can be easily combined, e.g. EMG sensors or AR HMDs. The usage of the widespread Unity 3D application environment enables high reusability and easy extension.

The following supplementary video to our research paper shows the system in action.

Controlling a real prosthesis
General setup, EMG pattern recognition (0:06), Two pre- and posttest trials (0:34 to 1:38)

Controlling a virtual prosthesis in augmented reality
General setup, EMG pattern recognition (1:39), Eight AR intervention trials (2:27 to 5:03)