ABSTRACT

In prosthetic arm control, to improve the robustness and reliability of different hand motions it is necessary for the arm to be controlled by highly accurate input or commands. To achieve high-accuracy input different types of classifier are used for pattern recognition. In this chapter the authors work with the widely used KNN (K-nearest neighbour) classifier and also determine time domain (TD) features. The output of the TD features works as input of the classifier. As the input of the classifier increases, the accuracy of the classifier also increases. In this chapter the authors work with six different hand gestures/motions for pattern recognition. For all hand movements, a power spectrum density graph is also significantly analysed. The proposed techniques achieved 87% of accuracy for different hand movements.