Unobtrusive Real-Time Monitoring of Pilot Mental Status: Development of a Test-Bed System

摘要

The real time determination of pilot mental and physical status is a critical feature of the workload monitoring and Mindware subsystems that have been envisioned for future jet aircraft. Recent laboratory and simulator studies, using retrospective data analyses, have suggested the value of various behavioral and physiological indices for reflecting task performance. The purpose of the present work was to develop software algorithms to derive some of these measures of interest in real time and to develop a test bed in which to explore the efficacy of these measures for inferring operationally relevant changes in pilot status. The present project demonstrated the feasibility and usefulness of this approach. Data processing algorithms were developed for characterizing. and integrating physiological indices based on heart rate and heart rate variability (vagal tone), eye blinks, and single trial, scalp recorded event related potentials. These physiological measures were obtained concurrently with behavioral measures as subjects performed a PC based, aviation simulation task. The data processing algorithms were implemented in a distributed processing configuration, using multiple personal computers, with the derived measures being integrated by a Decision Maker processor. This multi-processor test bed was demonstrated to work in near real time and attained encouraging levels of accuracy in characterizing the physiological phenomena of interest. Also encouraging were preliminary efforts to define decision rules, customized for individual subjects, that reflected a sensitivity of the measures derived in real time to manipulations of task difficulty.