Drunk driving is one of the most common reasons for road accidents. To tackle this issue, various solutions such as in-car breathalyzers are on the market. However, the breathalyzer method used in existing breath alcohol ignition interlock devices (BAIIDs) can easily be cheated. In this study, a web-based maze game was designed to detect the influence of alcohol in users, through recognizing their cognitive and motor skills. Additionally, the developed prototype was evaluated through a qualitative online survey. The survey results show the potential to replace existing devices with a mobile game, and point out possible directions to improve the current prototype so that it can be used as a valid ignition interlock device (IID) in the future. Due to law restrictions, the prototype could only be analyzed in a limited test environment with sober participants, resulting in no statistical evidence that the method used in this study could determine blood alcohol concentration (BAC) in a driver. Finally, eye-tracking was also integrated into a later version of the prototype, which was not evaluated in this study.

Contribution

This is a group project cooperated with 2 other KTH students. In this group, I was responsible for the integration of the eye-tracking technique, data analysis and the Latex typesetting of the final project. Besides, I jointly worked with another group member on the game development, in which I focused on the random maze generation and the implementation of different game states.