Scientific publication

Towards Real-time Probabilistic Evaluation of Situation Awareness from Human Gaze in Human-Robot Interaction

Publication from Digital

Paletta, Dr Lucas Paletta, Dini, A., Murko, C., Yahyanejad, S., Schwarz, M., Lodron, Gerald Lodron, Ladstätter, Stefan Ladstätter, Paar, Gerhard Paar, Velik, R.

Proc. Companion of the 12th Annual ACM/IEEE Conference on Human-Robot Interaction, HRI 2017, Vienna, March 6-9, 2017. , 1/2017


Human attention processes play a major role for optimization in human-robot interaction (HRI). This work describes a novel methodology to measure situation awareness in real-time from gaze interaction with scene objects of interest using eye tracking glasses and 3D gaze analysis. A probabilistic framework of uncertainty considers coping with measurement errors in eye and position tracking. Comprehensive experiments on HRI were conducted with tasks including handover in a lab based prototypical manufacturing environment. The methodology is proven to predict a standard measure of situation awareness (SAGAT) in real-time and will open new opportunities for human factors based performance optimization in HRI applications.