Properties Investigation of Surface Electromyography Signals using Centroids Estimated by Power Spectrum

  • : Universidade Federal de Juiz de Fora (UFJF)
  • : Surface electromyography signals (sEMG) has been applied in many fields of knoledge, such as biomechanics, sports, medicine and technology for develop of robotic prosthetics. This prosthetics use parameters extract from sEMG in a machine learning algorithm to classify different types of movements. However, some of this parameters extracted do not have a deep study. The objetice of this paper is to study the characteristics of sEMG signal using a centroid, in frequency domain, for six different movements of hand. The sEMG signals were obtained from Flexor Carpi Ulnaris and Extensor Carpi Radialis, Longus and Brevis, extrated from a public database. They were investigated through centroid, or barycenter, parameter, which was estimated from power spectrum and considering a sucessive sections windowing of the signals. The database is composed of two chanels, five subjects, six movements and thirty repetitions. The results were analyzed by visual inspections of the barycenter values, whith are grouped in histograms, and it shows how complicated is the subjects and movements distinctions. On the other hand, the results analyzed by using estatistic methods, medians comparison, with a confidence interval of 95%, shows there is possible a group distinction between movements and subjects.
  • : próteses
  • :