Using gaze point estimation and blink detection, to control robot arm by use of the EOG signal

摘要

Currently, there are many systems to support people with disability in their day to day activities. In this research, a robot arm system was designed, developed and controlled by use of an EOG (ElectroOculography) signal generated by the eye movement. The first part of the research was to establish the relationship between gaze position and EOG signal. Using Helmert transform method, the gaze position is estimated. The average error between real gaze position and estimated gaze position was ±4 cm. To recognize blink through EOG signal, Bell phenomenon was used. The blink recognition (voluntary and involuntary) experiment includes three occupants. The overall success rate of recognition of their blinks was more than 90%. Using this system long will make operators tired and results in erroneous experimental results. Thus, the switch from “control on” to “control off” was introduced in this system. The switch is activated when voluntary blink is detected. This will allow the system to improve the controllability.