Functional analysis of nonstationary EMG signals during a quick movement

An analysis is presented of the functional characteristics of the electromyogram (EMG) for a quick movement of the hand. The EMG signals resulting from a quick movement are represented by an autoregressive (AR) model. The AR parameters are estimated by the forward-backward model method, which is useful in the estimation of a nonstationary or dynamic signal. The prediction error variance, which is a peak frequency of EMG power spectrum, and the zero crossing rate are analyzed. The angular velocity of the hand during a quick movement is measured directly by placing an angular sensor on the joint axis of the wrist. Experimental results for five normal subjects are presented. They show the advantages of using the proposed method for nonstationary EMG signals