Analysis of chaotic characteristics on sEMG in hand gestures recognition

Since the existing feature extraction methods can not satisfy the demands from the controlling of myoelectric prosthetic hand for the sick of the stability and adaptability, there is still some distance for the disabled to use prosthetic flexibly. This paper analyzes the chaotic characteristics on surface electromyography (sEMG) of three hand actions. Three groups of sEMG were firstly collected from six muscles which played an important part of hand actions with strict experimental methods. And then, two kinds of chaotic characteristics which namely were fractal dimension and maximal Lyapunov exponent of three actions were extracted. Finally, we researched and discussed the clustering property of these two chaotic characteristics in detailed. The results indicate that the fractal dimension and maximal Lyapunov exponent of different hand actions can produce some certain clustering property, which means that the chaotic characteristics of sEMG can be used to identify the different hand gestures with some benefits on the stability and adaptability of myoelectric prosthetic hand.