Low-Cost Wireless Electrooculography Speller

• Published in: Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics pp. 123 - 128
• Main Authors: Ka Hou, Hoi, K.G., Smitha
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2018
• Subjects: Electrodes; Electroencephalography; Electrooculography; Electrooculography (EOG); EOG speller; human-computer interface (HCI); Keyboards; low cost signal processing system; Monitoring; Wireless communication
• ISSN: 2577-1655
• DOI: 10.1109/SMC.2018.00032

While communication is an innate ability to most of us, individuals with severe motor disability such as amyotrophic lateral sclerosis (ALS) have difficulty in daily communication. Nowadays, with the help of human-computer interface (HCI), they can communicate using bio-signal which can be measured from the human body and monitored. Electrooculography (EOG) is one type of bio-signal, it can be measured as the electrical potential difference generated by eye movement. This project aims to develop a low-cost wireless EOG based speller. The EOG signal acquisition used 5 surface electrodes, which captures horizontal and vertical eye movement. A low cost wireless EOG signal acquisition circuit was designed and implemented. A real-time EOG classifying algorithm was developed on the Arduino to classify up to 10 different types of eye movement. This classified information is transmitted wirelessly to a virtual keyboard ona computer. A self-design keyboard was developed by placing the most frequently used alphabets closer initial point to increase the efficiency of the speller. A test was conducted on eight subjects to compare the accuracy and efficiency of the proposed low cost wireless EOG speller. The proposed system was able to outperform most of the spellers reported in literature by providing an average accuracy of 89 % and speed of 6.48 letter per minute.