Intelligent invocation: Towards designing context-aware user assistance systems based on real-time eye tracking data analysis

Abstract

Recently introduced virtual and augmented reality devices such as the HTC Vive Pro Eye or Microsoft’s HoloLens 2 come with integrated eye tracking technology. Eye tracking technology is thus closer to the consumer market than ever before. Should these systems make the leap into the end consumer market, possibilities also evolve to use the data to feed intelligent user assistance systems in real time. One application could be to detect phases in consumers’ decision making processes based on eye tracking data, which, in turn, can be used to offer context-aware assistance to consumers. By analyzing eye tracking data from an experiment in a virtual reality shopping environment, we test existing approaches to detect decision phases and evaluate their applicability for an intelligent invocation of real-time user assistance. Furthermore, we propose a new approach, called on-the-fly-detection, since we conclude that existing approaches are not suitable for real-time phase detection.