Eye-Tracking Sensors for Adaptive Aerospace Human-Machine Interfaces and Interactions

Whereas current aerospace Human-Machine Interfaces and Interactions (HMI2) are mostly static in their behaviour/appearance and require direct input from human operators, innovative HMI
²
concepts are being proposed which allow for multimodal interactions or sense the functional state of human operators and dynamically adapt the level of automation. In particular, to facilitate trust between the human and machine, such systems need sensors that can reliably detect changes in the operator state. However, a number of environmental factors can affect the sensor's accuracy and precision. When used together with other novel sensors, eye tracking hasa significant potential to enhance the adaptiveness of aerospace HMI2. This paper presents the activities carried out to quantify the uncertainty associated with eye tracking equipment available in the Avionics and Air Traffic Management (ATM) systems laboratory of RMIT University, whichis being used to support the development of Cognitive HMI
²
. The presented methodology is used to characterise its measurement uncertainty based on a number of considerations, including the calibration error as well as gaze angle in static and dynamic conditions. The uncertainty associated with the eye tracker is used for error budgeting of a Cognitive HMI2 system which employs fuzzy logics to infer operator cognitive states based on eye tracking inputs.