A performance comparison of hand motion EMG classification

Abstract

Powered prosthesis is of considerable value to amputees to enable them to perform their daily-life activities with convenience. One of applicable control signals for controlling a powered prosthesis is the myoelectric signal. A number of commercial products have been developed that utilize myoelectric control for powered prostheses; however, the functionality of these devices is still insufficient to satisfy the needs of amputees. For the purpose of a comparison, several electromyogram classification methods have been studied in this paper. The performance criteria included not only classification accuracy, but also repeatability and robustness of the classifier, training time for online training performance, and computational time for real-time operation were evaluated with seven classification algorithms. The study included five different feature sets with time-domain feature values and autoregressive model coefficients. In summary, the quadratic discriminant analysis showed a remarkable performance in terms of high classification accuracy, high robustness, and low computational time of training and classification from the experiment results.

Research on Lower Limb Motion Recognition Based on Fusion of sEMG and Accelerometer Signals

“…Nowadays, many methods of feature extraction have been proposed, among which time-domain features and the wavelet coefficient are most commonly used [38][39][40]. The most commonly used time-domain features are Root Mean Square (RMS), Zero Crossing (ZC), Willison Amplitude (WAMP), SSC and Wave Length (WL).…”

Comparisons on different sEMG-features with dimension-reduction methods in hand motion recognition

“…In general, more features should have resulted in better accuracy. However, it was inferred that the mean accuracy result was not improved monotonically with increasing number of features [20]. As more features brought about more information relating to hand motion recognition, more noisy information of classification was enhanced and involved, too.…”

Pattern Recognition of Single-Channel sEMG Signal Using PCA and ANN Method to Classify Nine Hand Movements

“…The studied pattern recognition of the hand gesture is summarized in Table 1. The utilized number of channels in sEMG varies from two to 16 channels [6][7][8][9][10][12][13][14]. The selected sampling rate for the hand gesture recognition varied from 1 kHz to 4 kHz.…”

“…The studies presented in Table 1 show that the support vector machine (SVM), k-nearest neighbors (KNN), and artificial neural network (ANN) have been commonly used for EMG pattern recognition as classifiers. Most of the previous research employed sEMG signals with more than one channel [6][7][8][9][10][11][12][13][14]. The range of the sample frequency of the sEMG data acquisition was from 1 kHz to 4 kHz.…”