Real-Time Measurement and Analysis of Physiological Signals using Eye Tracking and Machine Learning

Important results from the project

Real-time measurement and analysis of physiological signals from the pilot will be useful to assess workload levels, fatigue, and support real-time attention guidance in the Man-Machine Interface. Recent research has shown that eye movements and pupil behaviours are related to multiple cognitive and affective states such as attention, mental workload, task difficulty, and task completion etc.

Expected long term effects

This project has investigated next-generation eye tracking and analytics for pilots with a special focus on between-fixation movements of the eye (i.e., saccades), micro movements and eye blinks related to the above discussed affective and cognitive states.

Approach and implementation

This project has developed and evaluate a head-worn demonstration system that measures and analyses the eye and the pupil to support attention and assess workload levels in operations. The system utilizes a Machine Learning (ML) approach for real time analysis together with low-power hardware inference engines. A suitable training set was collected in a simulator and processing of the data was done in a supercomputer.