Sistema de entrenamiento virtual inmersivo para prótesis robóticas mioeléctricas

Abstract

The development of myoelectric robotic prostheses has improved the autonomy of people with limb loss, although their high cost and the complexity of muscle control lead to high abandonment rates. In order to solve this problem, a virtual reality training system has been created for users with transradial amputation, allowing them to train their prosthetic control before acquiring the physical device. This system integrates EMG acquisition using the Noraxon miniDTS device, signal processing and classification in MATLAB, control through ROS 2 and a virtual environment in Unity that incorporates the RH8D hand model, visualized with an Oculus Quest. Validation with healthy subjects has shown improvements in execution time, good usability and a high potential as an efficient, low-cost alternative to conventional pre-prosthetic rehabilitation.