Implementation and testing of biofeedback-based strategies to improve bodily signals in virtual reality for myocontrol

Abstract

Disability takes many forms but among them the loss of a limb, especially the upperlimbs, is a source of major challenges for people in their everyday lives. So far there is not multifuncional prosthesis that is reliable enough to substitute a real arm and hand, even though the research on robotic prosthesis is extensive. Most of this research is focused on the final stages of learning and predicting, but there is little research on the actual gathering of data and if it influences the outcome of said learning and prediction stages. Also, in the traditional training processes for robotics, the user is rather a passive element that repeats a given procedure rather than an active participant that can influence the whole process. For all these reasons, this research project focuses on developing a game-like environment that measures the quality of the data generated by the user, using metrics such as the inter and intraclass variability of the clusters of actions, and, at the same time, engages the user to actively take part of the process. Initial results done in a few participants, show that the approach is promising although it can not be sufficiently proved by this project.