Brainwave-based and motion-based human computer interface for disabled people

Abstract

Computers are playing an important role in daily life. However disabled people can fnd it diffcult to use computers because of the diffculties in controlling mouses and computer keyboards. They need a special human computer interface (HCI) to facilitate them in using computers. This research project is divided into two parts. The frst part is about brainwave-based HCI technology. Using a brain-computer interface, paralyzed users can control computers more effectively. This project seeks to explore using brainwave technology to assist disabled people in using computers. A low cost brainwave sensor was adopted to develop a classifer program for classifying binary input for disabled people. Three classifcation methods, namely neural network, support vector machine and boosted decision tree were investigated. It was found that the boosted decision tree method achieved the best performance. By using the boosted decision tree method, the classifer program could successfully classify brainwave data with an accuracy of 80 percent. The second part is about motion-based HCI technology and a newly designed virtual keyboard was investigated. The aim is to minimize the typing cost for disabled users. A genetic algorithm-based keyboard was designed that could reduce the typing cost by about 40% as compared with a normal keyboard. A virtual keyboard prototype was developed based on the Leap Motion Controller. Furthermore, motion technology is helpful for disabled people to do rehabilitation exercises. A motion-based training game for occupational therapy was developed in collaboration with some students. By combining motion sensor technology with a digital game, disabled people can perform hand training exercises more effectively.