Review of data-driven and learning-based fault detection and diagnosis in satellites: Individual and formation configurations

With the growing deployment of both single satellites and formation-flying (FF) satellite systems in advanced space missions, the reliability and safety of attitude control systems (ACS) have become critically important. In this context, the rapid advancement of data-driven and learning-based approaches have emerged as a powerful alternative to traditional model-based methods for fault detection and diagnosis (FDD) in aerospace systems. This paper provides a review of data-driven FDD methodologies, specifically machine learning techniques applied in both configurations. Initially, the study investigates data-driven FDD mechanisms for individual satellites, focusing on machine learning algorithms. A structured comparative assessment follows, detailing the advantages and limitations of each methodology. The scope is then expanded to address data-driven FDD in FF scenarios. Through a comprehensive analysis, this review identifies emerging research trends and significant methodological overlaps within the domain.