Removing Artifacts from EMG Envelope for Improving Myoelectric Robot Arm Control

There are analysis methods of biomedical signalfeatures at present to know any information regarding the humanbody. It is to use the extracted features from the EMG signalto predict human motion and its associated efforts by usingsignals given by the motor unit action. Steps to process theEMG signal are envelope acquiring, artifacts filtering, estimationsmoothing, EMG value standardizing, feature classifying, andmotion recognizing. Different methods are useful to achieve thisgoal and apply by experimental projects. Using a database ofEMG signals, we calculate the envelope by using the rectifiedsignal, where we take the absolute number of EMG signals so thatall values become positive. In the first step, we shall now proceedto remove EMG envelope artifacts by using filters such as theKalman filter (KF), H1 filter, unbiased finite impulse response (UFIR), cKF, cH1, and cUFIR. The last three algorithms wereamended by assuming colored measurement noise. Last, we makea standardization of the EMG envelope. Given the above, we willknow if the estimation envelope gives the optimal features for anaccurate prediction.