We propose to develop and demonstrate an artificial intelligence / machine learning (ML) algorithm which will reveal insights into astronaut health and performance, which in turn can be used to reduce mission risk, and optimize spacecraft resources and limited astronaut time during the mission. The proposed ML algorithm can be applied to any astronaut biometric data including heart rate, blood oxygen, temperature, nutrition in consumed food, environmental control system parameters, and injury and illness reports to better inform astronauts and the Mission Control Center with actionable insights made available in real time. Patterns in these data can indicate that astronauts are not at peak physical or mental performance, and patterns in these data sets can suggest how to return the astronauts to peak performance.
This algorithm can provide critical insights to astronauts by recognizing subtle patterns that appear across many telemetry channels. For example, this algorithm could track metabolic activity in crew in real time and suggest changes to environmental system parameters to maintain constant CO2 levels in a most resource-efficient way. Another example is early warning of dangerous conditions (precursors to hypercapnia). Its diagnostic power lies in its ability to learn in real time, and also in its ability to find subtle correlations between multiple bio parameters in data telemetry available to it. This algorithm also has the added benefit of fast learning for adapting to nuances between astronauts' metabolic profiles and finding new diagnostic correlations in real time during the mission.
As NASA sends crew out to the Moon and beyond to deep space and Mars, the lack of a quick return to Earth means early diagnosis of physical and mental health problems could mean the difference between life and death. This algorithm is compatible with low power hardware so it can be performed onboard spacecraft in deep space where there is limited data connectivity to Earth.
This product is directly applicable to future crewed NASA missions that are at the Moon or beyond where there is less opportunity for a quick return to Earth. In these cases biomarker monitoring can provide early warning for health concerns needing attention. In these cases also resources are even more precious and the product can help optimize use of those limited resources. Our proposed innovation works with any time series data available, such as that provided by the International Space Station (ISS) Food Intake Tracker (ISS FIT) iPad App.
Personalized monitoring and tracking precursors for optimal performance and health applied to the following markets: