10th International Conference on Augmented Cognition (AC) Held as Part of 18th International Conference on Human-Computer Interaction (HIC International), Toronto, Canada, 17 - 22 July 2016, vol.9743, pp.147-158
Real-time monitoring of the flight crew's health status with ambient and body sensors have become an important concern to improve the safety and the efficiency of flight operations. In this paper we report our preliminary findings on a functional near-infrared spectroscopy (fNIR) based online algorithm developed for real-time monitoring of mental workload of an airline pilot. We developed a linear discriminant analysis (LDA) based classifier that aims to predict low, moderate and high mental workload states based on a set of features computed over a moving window of oxy-and deoxy-hemoglobin measures obtained from 16 locations distributed over the prefrontal cortex. In this paper we explore the predictive power of a model trained for a single pilot over a sample of eight pilots and discuss the technical challenges involved with real-time measurement of brain activity in a flight simulator environment that involves other infra-red sources.