Reducing Aviation Fatalities by Monitoring Pilots' Cognitive States Using Psychophysiological Measurements
Abstract
Airplane accidents are usually catastrophic, and the majority of flight-related accidents are caused by a lack of situational awareness during flight. To improve flight safety, we built a model to detect the cognitive states of pilots from their psychophysiological signals so that the aviators can be warned before falling into a dangerous mental state, including channelized attention, diverted attention, and startle/surprise. The research is composed of time series analysis and classification. We used seasonal decomposition, exponential smoothing, and autoregressive integrated moving average models to analyze the numerical psychophysiological measurements of 18 pilots and utilize such measurements to distinguish their cognitive states by classification methods, such as random forest, support vector machine, and logistic regression. The results can be a part of the risk management mechanism to alert pilots when necessary. The deliverables include a classification model of the problem and an analysis of the solutions obtained from the model. These models are written in R so that anyone can run calculations in real time to monitor the cognitive states of pilots and to support follow-on/future analysis work.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2021
- Accession Number
- AD1151053
Entities
People
- Yi-Chung Lin
Organizations
- Naval Postgraduate School