Hardware and software implementation of real time electrooculogram (EOG) acquisition system to control computer cursor with eyeball movement

Abstract

Human computer interface (HCI) is an emerging technology of neuroscience and artificial intelligence. Development of HCI system using bio signal e.g. Electrooculogram (EOG), Electromyogram (EMG), Electroencephalogram (EEG), Functional near-infrared spectroscopy (fNIRS) etc. are attracted more and more attention of researchers all over the world in recent years because through this it is possible to get acquainted with advanced technologies of artificial intelligence. This paper presents the design and implementation of a fully functional Electrooculogram (EOG) based human computer interface. In this work we have designed and implemented necessary hardware and software for EOG signal acquisition along with controlling hardware such as wheelchair, robotic arm, mobile robot etc., and move computer mouse cursor simultaneously using EOG signal. This interface has three portion: EOG signal acquisition and amplification, analog to digital conversion, and real time hardware and mouse cursor movement. Eye movement is detected by measuring potential difference between cornea and retina using five Ag-Agcl disposable electrodes. Frequency range of EOG signal is considered as 0.3 to 15Hz, so this frequency range is taken using an active high and low pass filter so that accurate EOG signal can be achieved. The analog output of the EOG signal from filter is converted into digital signal by using an Arduino. Arduino serialize the EOG data for calibration and provides a threshold reference point which is used for controlling Hardware. The Classification module e.g. Support Vector machine (SVM) and Linear Discriminant Analysis (LDA) classify live data with respect to the horizontal and vertical data. This works as a binary classifier and choose optimal hyper-plane between two variables. According to each update on the eye position, cursor automatically accelerated in particular direction. PyMouse module in python is used for this task. Eye gesture based Hardware like robot, wheelchair etc. control and mouse cursor movement are the principle outcome of this research work.