Finger movements classification from grasping spherical objects with surface electromyography using time domain based features

Guadalupe A. Torres, Victor H. Benitez

Resultado de la investigación: Contribución a una conferenciaArtículo

5 Citas (Scopus)

Resumen

© 2016 IEEE. In this paper the classification of fingers gestures that vary in specific mechanical positions is proposed, which consist in distinguish several finger positions with very low mechanical variation. A new approach is presented that is different to the state of the art methods for the classification of fingers movements that have traditionally, focused on very well distinguished gestures from each other. Myoelectric signals (MES) reflect the intention of the movement according to the diameter of the sphere sustained is the objective of the present study. Natural motions are collected by placing electrodes on five muscles on the forearm of six healthy subjects, while performing spherical fastenings. A time domain (TD) feature vector is given as inputs to a linear discriminant analysis (LDA) module. LDA is used as statistical pattern classifier. We show that there exist significant relationship between muscle signals and fingers positions. Therefore, it is possible to categorize each class of finger position, that is, TD feature based provide an effective representation for classification. LDA achieve the assignment of the membership of a MES collected to one fingers position class, which is defined by the diameter of sphere. These results will be useful for analysis of movement of the human hand to improve control of robotic prosthetic hand and man-machine interfaces.
Idioma originalInglés estadounidense
DOI
EstadoPublicada - 28 nov 2016
Evento2016 IEEE Conference on Mechatronics, Adaptive and Intelligent Systems, MAIS 2016 -
Duración: 28 nov 2016 → …

Conferencia

Conferencia2016 IEEE Conference on Mechatronics, Adaptive and Intelligent Systems, MAIS 2016
Período28/11/16 → …

Huella

Profundice en los temas de investigación de 'Finger movements classification from grasping spherical objects with surface electromyography using time domain based features'. En conjunto forman una huella única.

Citar esto