In Slovenia, eye trackers are still very much unknown. Undoubtedly, one of the main reasons lies in the fact that this technology has not yet reached the final user, although the beginnings of eye tracking go back more than 150 years, to the end of the 19th century, when first automobiles were produced. Yet, these days almost every family owns at least one car, but the same cannot be claimed for eye tracking devices. This is surprising considering that since its beginnings the mankind is agitated to discover, how would it be to live in the world, in which we could control things around us using just eyes. After all we are living in an era of new devices entering the market with their prices being lower than the price of an average mobile phone. Could this finally be the reason to launch a mass usage of this technology, which is still unknown to the majority of people?
The first part of this master's thesis focuses on the history of eye trackers - since the beginning of the production and the first attempts trying to record eye movements, to the development of the trackers and their current form. The chronology itself includes some eye characteristics, discovered by those first eye trackers. The subsequent chapter discusses other eye characteristics which play an important role in the working principle of eye tracking and need to be explained in order to understand the capabilities of eye trackers. The chapter is followed by the presentation of four most commonly used eye tracking techniques, their operating principles and the applicability, as well as their capabilities.
Later on the master's thesis presents the problems of eye trackers being used as communication channels for entering information into a selected system through controlled eye movements. Since the human eye is primarily designed for viewing and not for giving orders, i.e. controlled eye movements, this chapter describes a typical deficiency of an eye – its precision of indicating can be compared to the thickness of a finger and it is impossible for an eye to confirm an action in an unforced manner – with a click. Consequently, the thesis emphasizes the need to adjust the functioning of operating systems and applications in accordance with »non-command based system«, manipulated with a series of continuous events which is closer to the natural behavior of an eye. However, since currently no such devices exist on the market, the thesis describes some of the most common methods of adjusting eye tracking interface in order to function in modern applications and operating systems.
Still, eye trackers were not designed solely as an alternative input channel for a mouse and keyboard when working with a computer. At the beginning eye trackers were designed as a tool for investigating eye movements, but close connection with cognitive processes in humans was quickly discovered. Therefore, scientists soon started to use eye trackers to study cognitive processes. The practical part of the master's thesis includes the integration of an eye tracker The Eye Tribe into a driving simulator in order to enable a statistical analysis of a driver's behavior and his cognitive load in simulated traffic situations. Since this is the first time a low price eye tracker was used in this manner, I wanted to investigate its practical aspect in collaboration with LaIT laboratory. We have defined investigatory qualifiers regarding the reliability of operating, accuracy and precision, the size of an area of interest (AOI) the device is still capable to follow, as well as the possibilities of detecting eyesight narrowing (tunnel vision) and changes in pupil dilation when the driver is under greater cognitive load.
To this end, we have drawn up reference measurements. Measurements were performed in accordance with the draft standard ISO 17488. Within the scope of this standard we have used Delayed Digit Recall in order to achieve different levels of cognitive load and tactile DRT method to indirectly measure cognitive load on the basis of reaction times.
According to the study and compared to the DRT method the low-price The Eye Tribe device proved to be a successful tool for indicating cognitive stress of a driver. It is important to emphasize that this is not an invasive method and it does not additionally affect the cognitive load of a driver. Furthermore, the method does not affect the primary task which is the subject of testing.