A study on the motion decision of the arm using pattern recognition of EMG signal

EMG신호의 패턴인식을 이용한 동작판정에 관한 연구

In this paper, the primitive and double combined motion classification of the arm is discussed using pattern recognition of EM signal. The EM signals are detected from Ag-Ag/Cl surface electrodes, and IBM PC, calculated the Likelyhood probability and the decision function on the feature space of integral absolute value. Multiclass decision rule is introduced for higher decision rate. On our experimental results from expert simulator, the decision rate of more than 78% can be obtained.