Journals Proceedings

Abstract

Direct Human Input (DHI) is an input methodology for Human Machine Interfaces (HMI). Use of DHI eliminates physical input devices as input data is collected directly from the signals generated by the human body. Various varieties of DHI exists of which Electromyogram (EMG) is found to have properties that makes it readily applicable. In this work, the capability of EMG to detect facial expressions of the user has been utilized to create a generalized interface that processes EMG signals into machine commands in accordance with the user’s facial expressions. The interface is essentially a combination of dimension reduction and classification algorithms that classifies EMG signals into facial expressions to activate machine commands correlated with the estimated facial expression. Linear Discriminant Analysis and Support Vector Machines have been used for dimension reduction and classification respectively. The interface has been found to perform satisfactorily with a classification accuracy over 99%. Case studies were conducted to verify the performance of the proposed interface in real-time applications and the performance was found to be satisfactory.