Finger movements recognition using minimally redundant features of wavelet denoised EMG

摘要: Developing prosthetic hands with high functionality and ease of use is the focus current research in area Electromyogram (EMG) based prosthesis control. Although individuals upper limb loss can perform grasping operations currently available hands, more intuitive control finger movements required to replicate complex motor functions human hands. A significant challenge classify higher recognition rates using a smaller number EMG channels. This paper reports novel criterion for selection minimally redundant feature set classification 10-class two-channel EMG. The selected from wavelet denoised at four decomposition levels minimum redundancy terms mutual information. five features: root mean square, simple square integral, slope sign change, peak frequency power spectral ratio have been 31 time domain features. Using state art technique on support vector machine, we achieved 10-fold cross-validation rate 96.5 ± 0.13%. experimental study shows that information ensures highest reduced computational cost.