Human Error Prediction Using Eye Tracking to Improvise Team Cohesion in Human-Machine Teams

Abstract

The rapid increase in integration of intelligent systems in every corner of technology created an emerging field called Human-Machine Teaming (HMT). In HMT, human and machine collaborate with one another to accomplish a common goal or task. To achieve the best performance of a team, it is necessary to build trust and cohesion among all teammates (machines and humans). Furthermore, in a team, it is an established fact that a team member ability to predict fellow member’s future course of action and have an accurate picture is a valuable asset and will result in better team dynamics and team performance. To realize such an ability we are proposing a human error predicting methodology that could give an intelligent system a better understanding of human actions in advance. In the proposed method, we used eye-tracking metrics such as gaze density and cognitive state to predict human errors. The results obtained with the proposed and developed methods are found to be efficient in predicting human error probability.