Recognition of Movements Through Dynamic Electromyographic Signals

FARFÁN, FERNANDO DANIEL; SOLETTA, JORGE; RUIZ, GABRIEL; FELICE CARMELO JOSÉ

International Journal of Engineering Research & Technology

IJERT Editorial Management System

Lugar: Bombay; Año: 2016 vol. 5 p. 450 - 457

2278-0181

The recognition of movements throughelectromyographic (EMG) signals is critical for myoelectriccontrol systems. Performance of these systems depend onprocessing methods and protocols used to extract the EMGsignals. The aim of this study is to evaluate the performance ofclassification of a kinematic recognition system based ondynamic EMG signals. For this, a correlation analysis betweendynamic EMG signals and kinematic features of movements isrealized, and then, a kinematic recognition system based ondynamic EMG signals is implemented. Dynamic EMG signalsfrom forearm muscles during finger flexion movements wererecorded and analyzed by using an amplitude estimator. Linearand no-linear correlations between EMG amplitudes andkinematic features were found. Then, a step of classificationbased on discriminant analysis was implemented to categorizethe finger movements in multiple kinematic states. The accuracyof classifications were 95%, 88%, 81% and 76% for two, three,four and five states respectively, and by using a simple-channelrecording and an EMG amplitude estimator. The results of thisstudy demonstrate that it is possible to improve aspects of?intuitiveness? through dynamic EMG evoked by natural andmore intuitive movements.