Classification for movement identification from EMG signals

Laura Daniela Lasso Arciniegas and Brayan Andres Viveros Melo,Universidad de Nariño, San Juan de Pasto, Colombia-2019

Electromyographic signals are electrical impulses produced by the muscles during the processes of contraction and relaxation, this feature allows analyzing the physiological state of the muscles and also allows creating control interfaces for communication between user and machine. Among the different applications of interfaces, those dedicated to individuals with motor disabilities that impede the performance of daily activities have taken off. Therefore it is pertinent to perform a search of characteristics and evaluation of machine learning techniques, to define criteria that allow carrying out this type of control interfaces.

Electromyographic signal.

Figure 1. Example of an electromyographic signal.

In order to carry out a classification task, it is necessary to carry out an adequate segmentation, characterization, dimension reduction and finally classification procedure. Finally, once a motor intention order is classified, it must be executed by an actuator of interest.

Figure 2. View of the interface implemented on python supported with PyQt Designer.

GUI- Prototype hand prothesis training.



Exploration of Characterization and Classification Techniques for Movement Identification from EMG Signals: Preliminary Results.
Colombian Conference on Computing 2018.
A. Viveros-MeloL. Lasso-Arciniegas. J. A. Salazar-CastroD. H. Peluffo-OrdóñezM. A. BecerraA. E. Castro-OspinaE. J. Revelo-Fuelagán.

Movement Identification in EMG Signals Using Machine Learning: A Comparative Study
International Workshop on Artificial Intelligence and Pattern Recognition 2018.
Laura Lasso-Arciniegas. Andres Viveros-MeloJosé A. Salazar-CastroMiguel A. BecerraAndrés Eduardo Castro-OspinaE. Javier Revelo-FuelagánDiego H. Peluffo-Ordóñez

About the authors.

 Andres Viveros Laura Lasso
 Electronic Engineer Electronic Engineer