SURFACE EMG CLASSIFICATION USING MOVING APPROXIMATE ENTROPY AND FUZZY LOGIC FOR PROSTHESIS CONTROL

Abstract

Electromyographic control systems based on pattern recognition have become an established technique in upper limb prosthetic control application. This paper describes a use of fuzzy logic to discriminate different hand grip postures by processing the surface EMG from wrist muscles. A moving data window of two hundred values is applied to the SEMG data and a new method called moving approximate entropy is used to extract information from the signals. The analyses show differences at three states of contraction (start, middle and end) where significant dips can be observed at the start and end of a muscle contraction. Mean absolute value (MAV) and kurtosis are also used in the extraction process to increase the performance of the system. The extracted features are fed to a fuzzy logic system to be classified and select the output appropriately. The preliminary experimental result demonstrates the ability of the system to classify the features related to different grip postures.