A k-Nearest Neighbor Based Algorithm for Human Arm Movements Recognition Using EMG Signals

Abstract

In a human–robot interface, the prediction ofmotion, which is based on context information of a task,has the potential to improve the robustness and reliabilityof motion classification to control human-assistingmanipulators. The electromyography (EMG) signals canbe used as a control source of artificial arm after it hasbeen processed. The objective of this work is to achievebetter classification with multiple parameters using KNearestNeighbor for different movements of a prostheticarm. A K- Nearest Neighbor (K-NN) rule is one of thesimplest and the most important methods in patternrecognition. The proposed structure is simulated usingMATLAB Ver. R2009a, and satisfied results are obtainedby comparing with conventional method of recognitionusing Artificial Neural Network(ANN), that explains theability of the proposed structure to recognize themovements of human arm based EMG signals. Resultsshow the proposed technique achieved a uniformly goodperformance with respect to ANN in term of time which isimportant in recognition systems, better accuracy inrecognition when applied to lower SNR signal .