We have machine learning and crew state experience and expertise. We pioneered the study and use of real-time-compatible methods ten years ago. The best thus far used autoencoding to deal with biological noise, and two 6-layer dense neural networks. Non-nominal attentional states are identified at a rates > 0.8.


LAR-18996, US Patent 10,192,173. https://technology.nasa.gov/ patent/LAR-TOPS-88.

Selected Publications

Terwilliger, P., Sarle, J., Walker, S., Harrivel, A. A ResNet Autoencoder Approach for Time Series Classification of Cognitive State, MODSIM World 2020, Paper No. 0053, (virtual event).

Napoli, N., Stephens, C., Kennedy, K., Barnes, L., Juarez Garcia, E., Harrivel, A. NAPS Fusion: A framework to overcome experimental data limitations to predict human performance and cognitive task outcomes. Information Fusion. 2023; 91:15-30.

Harrivel, A., Weissman, D., Noll, D., Huppert, T., and Peltier, S. (2016). Dynamic filtering improves attentional state prediction with fNIRS. Biomedical Optics Express. 7(3), 979-1002.

Technical Points of Contact

Dr. Angela Harrivel, Crew Systems and Aviation Operations Branch, NASA Langley Research Center

Return to “Research Spotlight: NASA 2024 Human Research Program Investigators’ Workshop