AIR TOUCH: Human Machine Interface Using Electromyography Signals

Abstract Novel interactions between futuristic devices and humans in the ever-expanding digital world is gaining momentum in the current era. In this paper, a system is proposed where electromyography (EMG) signals are used to control the cursor on a PC with the movement of the hand, making effortless interaction between user and the computer. The hand movements are detected using accelerometer and EMG signals acquired using electrodes are used to classify the hand gestures. Time domain features are extracted from the EMG signals and the gestures are classified using K-Nearest Neighbor (KNN) classifier. The operation to be performed on PC is determined from the gesture with help of a suitable interface. This system is implemented to perform the positioning of the cursor and two of the most common actions of a mouse, namely, single click and double click. The system showed an accuracy of 98% in classifying the gestures.