IMPROVEMENT OF MYOELECTRIC PATTERN CLASSIFICATION RATE WITH µ-LAW QUANTIZATION.

Isamu Kajitani, Nobuyuki Otsu, Tetsuya Higuchi

Abstract

In order to realize a myoelectric-controlled multi-functional hand prosthesis, this paper proposes a method to improve the myoelectric pattern classification ability of a hand controller. By applying the proposed method of µ-LAW quantization, the pattern classification rate increased by 11.1% (averaged for five subjects) and by 15.5% (maximum), with a practical pattern classification rate of 97.8% being achieved.

Keywords

myoelectric, prosthesis, logic circuit

Download

PWC-2003-TC18-002.pdf

DOI

-

IMEKO TC

TC18 - Measurement of Human Functions

Back to the proceedings