In this thesis, several key aspects in the design of a commercially viable electrooculography (EOG) acquisition device are discussed. The context for the needs of such a device, along with associated background in the EOG signal and previous work are presented. The topic of electrooculography electrodes is introduced with a thorough investigation into seven different types, all chosen for their relatively small size and practicality. This investigation includes the impedance settling time, sweep response, Warburg model parameters and motion artifact data for each electrode, providing an objective overview. EOG signal variability from subject to subject is explored with strong correlation found. The majority of the face area is assessed for levels of electrooculography signals present with all areas showing potentially useful levels. Finally, a novel pattern recognition algorithm is used to assess the accuracy of EOG signal classification while exploring the effect of the number of electrodes used and their positioning.
Zircon - This is a contributing Drupal Theme Design by WeebPal.
