2014 | OriginalPaper | Buchkapitel
Study on Pattern Recognition of Hand Motion Modes Based on Wavelet Packet and SVM
verfasst von : Fuxin Liang, Chuanjiang Li, Yunling Gao, Chongming Zhang, Jiajia Chen
Erschienen in: Computational Intelligence, Networked Systems and Their Applications
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
For pattern recognition-based myoelectric prosthetic hand control, high accuracy of multiple discriminated hand motions is presented in related literature. But in practical applications of myoelectric control, considering cost and simple installation, fewer sensors are expected to be used. A method of pattern recognition based on the wavelet packet decomposition and support vector machine (SVM) is proposed in this paper. Firstly, energy spectrum as feature vectors of the surface electromyography (sEMG) signal is acquired by wavelet packet transform. Then, SVM is used for pattern recognition of hand motion modes. Four channels of sEMG signals obtained from sensors placed on different positions of forearm are used to experiment of hand motion recognition. And different combinations of 2 or 3 signals are tried to recognize hand motion modes. The results show that recognition rate of proposed method can get 92.5% while using 4 sEMG signals to recognize 8 different hand motions, which is 2.5% higher than using traditional method. And when using 3 sEMG signals from specific positions, it can reaches as high as 90%. When using 2sEMG signals only 6 motions can be discriminated with more than 90% recognition rate. Thus, the proposed method can meet the demands of sEMG prosthetic hand control and has high practical value.