An efficient handwriting velocity modelling for electromyographic signals reconstruction using Radial Basis Function neural networks

The detection and reconstruction of electromyographic signals become a major issue in the biomedical field since they are a necessary source of information in many clinical and industrial applications. People with disabilities or suffering from various neurological diseases are facing so many difficulties resulting from problems located at the muscle stimuli (ElectroMyoGraphic EMG signals) or signals from the brain (ElectroEncephaloGraphic EEG signals) and which arise at the stage of writing. Therefore, it is interesting to use an experimental recordings database in order to elaborate a model able to reconstruct the electromyographic signals relative to the handwriting production for different writers. This paper proposes a new neural approach for modelling and characterization of the handwriting process allowing the reconstitution of EMG signals from the handwriting velocities based on the exploitation of artificial neural networks concepts and more specifically the Radial Basis Function (RBF) neural networks. Our findings show a satisfactory agreement between the responses of the developed neural model and the experimental data for various letters and forms which demonstrates the efficiency of the proposed approach.