Evaluation of Electromyogram Signals in Athletes using Elman's Recurrent CNN

In this research, the electromyogram signals recorded from athletes using the Elman’s CNN were investigated. Among the most important cases presented, the proper use of desirable filters for better recording of EMG and of course methods to deal with unwanted noise are presented so that the noise combined with the signal and the noise caused by low-frequency motion artifacts are removed and minimized to a desirable level. A Butterworth filter with a cut-off frequencies of ۱۵ to ۱۰۰ Hz was used, and it should be noted that to obtain the desired result while a low-pass analog filter was used to record EMG with an anti-augmentation needle electrode. Also, using the Elman’s CNN, we have achieved a high classification accuracy of ۹۷.۹%, which indicates the relatively high accuracy obtained from the EMG signals recorded in athletes. In addition, the process of EMG signal processing, electrical characteristics of the mentioned signals, important parameters of vital signal recording, classification and analysis of signals, use of amplitude and linear phase as well as bandwidth and EMG amplifiers are covered, respectively

Keywords:

Electromyogram , Elman CNN , EMG Signals , Athletes , Sports