A preliminary investigation of selection of eeg and psychophysiological features for classifying pilot workload

摘要

Cardiac and respiration measures have shown limited success in classifying pilot workload. In the last few years, electroencephalography (EEG) measures have also been used to classify pilot workload (Sterman et al, 1992, and Wilson et al, 1994). Features of EEG data are currently selected by minimization of pilot workload classification error by a multilayer perceptron (MLP) classifier (Wilson et al, 1994). A set of salient features has not yet been idnetified for classifying pilot workload.