Life Support Systems: A Comparative Study of Machine Learning Applications, Challenges, and Future Developments in Space Exploration

Future space exploration missions will take humans farther away from Earth than ever before. This increases the necessity for more reliable and efficient life support systems (LSSs). To improve autonomy and safety, the use of modern machine learning (ML) methods for fault detection or control is promising. This paper provides a comparative study of existing research on ML applications, including technology from other scientific fields such as the automotive or smart home sectors, to motivate further research towards developing reliable and efficient ML-based solutions for LSSs in human space exploration. Challenges about explainability, robustness and overall implementation in the harsh environment of space must be addressed to ensure the safety of the crew and flexibility as well as scalability of a model. Looking ahead, key opportunities and possible future research in the areas of ressource optimization, autonomy and efficiency are discussed. In addition, the current and ongoing research in the field of LSSs at the University of Stuttgart’s Institute of Space Systems is highlighted, including an overview of the Human Spaceflight Lab and ML-based fault detection.