Regulation Stability Analysis Intelligent Learning - ARTEMIS Experiment SpacePatch. Autarkic Tiny AI Systems for Stability Analysis of cardiovascular Regulatory Mechanisms using intraindividual Time Series Feature Aggregation and longitudinal Machine Learning. (RESIL-ARES 50WB2321)
Third party funded individual grant
Acronym: RESIL-ARES 50WB2321
Start date : 01.12.2023
End date : 31.05.2025
Project details
Scientific Abstract
The-ARES "ARTEMIS Experiment Space Patch" project results on a "DLR Call for Ideas - AI for Space Applications".
The project combines high end wearable biosensors - e.g., the seismocardiography sensor for ARTEMIS II
/ ORION - with Tiny-AI based signal and time series processing for individualized Human Machine
Communication (HMC) with the goal to better understand and classify the mechanisms and stability of
cardiovascular regulatory processes (RES).
In the field of integrated performance physiology under extreme environmental and stress conditions, the
assessment and possible prediction of individual exercise capacity is essential. Intra-individual RES
analysis of potential risks that could lead to mission abort are core to the project.
The application of new AI learning models (IL) to the known but highly complex interplay of the regulatory
systems of different levels of the cardiovascular system and its mechanical and electrical biosignals shall
contribute to the "safe return" of the astronauts. This is the focus of the project, based on the novel
aggregation of time series features suitable for longitudinal RES learning (RESIL) - also under long-term
flight conditions, e.g., Tiny AI learning without an instructor.
In the project, components of a Tiny-AI with databases and interfaces will be developed, adapted and
aggregated in a modular, self-sufficient (internet and cloud independent), and suitable way for "ARTEMIS
and beyond" missions as a prototype system called RESIL-ARES. Applications will be derived from this,
e.g. the inflight system for ARTEMIS II and ":envihab" system for preflight learning and postflight analyses.
The self-sufficient, clinic-network-independent system to be set up by the applicant in the project comprises
the AI-eco-, development and test system for the new methods, as well as for their programming, the
evaluation system, and other necessary traditional AI components of biomedical and data sciences.