The Research Of Processing Technology For Surface Electromyography Signals

The skeletal muscle will make the following activities when the human haveactions. The corresponding bio-electric signal will be produced on the skinsurface. The bio-electric signal is a natural human action language and it reflectsthe functional status of body’s neuromuscular at the time of its activity. Hence,the experts in different fields have great interests in this kind of signal and carryout a wide range of research expecting to obtain some information about humanbody’s muscle tissue from the EMGs. The applications including control of theclinical diagnose, the rehabilitation engineering, the mechanical equipment andthe sport research are actively promoted by a large number of researches.The surface EMGs is a kind of weak bio-electric signal and it is easy to bedisturbed in the acquisition process. That makes the weak signal more difficult tobe recognized. Hence, the application of surface EMGs is limited in the field ofEMG equipment. This paper studies the pretreatment technology of surfaceEMGs and also analyzes the pretreatment technology suitable for the weak bio-electric signal as the surface EMGs. The adopted surface EMGs is made of byfour groups of surface electrode collection which are stuck on the body’s upperlimb, including muscles of the extensor digitorum, the flexor digitorum prefunds,the flexor carpi ulnaris and the skin surface of brachioradialis.This paper firstly analysises the generation mechanism of the surface EMGs,finds out its features and provides the reliable theory basis for the subsequentpreprocessing technology. Afterwards the noise of the surface EMGs by applyingthe wavelet threshold de-noising method is eliminated and a better signal on itsnoise ratio is extracted. Then the features of surface EMGs including the time-domain characteristics of the mean, the standard deviation, the mean square root,the frequency domain features of the power spectrum, and the time-frequencycharacteristics of wavelet coefficients are extracted. Finally, the characteristic clustering is analyzed and the more powerful characterization of surface EMGfeatures is judged. It provides the surface EMGs applied to the actual EMGequipment a reliable theoretical basis.