BioGuitar

The project group BioGuitar has tasked itself with creating a functional air guitar that produces sounds through movements. Within the project itself, there are three groups focusing on hardware, machine learning and synthesis.

Using biosensors, the user’s hand movements are being measured in realtime while they’re playing the air guitar. Through the use of machine learning models, these movements can be classified. Said information is then translated into MIDI data, from which sound is generated. The air guitarist can now feel like rock star from the comfort of their own home!

The movement is being measured with the help of a microcontroller, which posesses EMG sensors (electromyography) and accelerometers. In this way, hand grips and movements can be recorded uninvasively. This movement data is analysed and classified by previously trained ML models. From this classified data, the position and movement of the hands can be clearly determined and processed. For sound generation, standards such as MIDI for communication and the VST interface are used. This makes the project easily extendable and can be used for more purposes than a simple guitar.