Subject-independent modeling of sEMG signals for the motion of a single robot joint through GMM Modelization

This thesis evaluates the use of a probabilistic model, the Gaussian Mixture Model (GMM), trained through Electromyography (EMG) signals to estimate the bending angle of a single human joint. The GMM is created from the EMG signals collected by different people and the goal is to create a general model based on the data of different subjects. The model is then tested on new, unseen data. The goodness of the estimated data is evaluated by means of Normalized Mean Square Error