A Distributed Fault Detection and Estimation for Formation of Clusters of Small Satellites

This paper explores the problem of distributed fault detection and estimation for clusters of satellites. An observer implemented on each satellite can detect faults and estimate their size and behavior over time. Satellite observers can monitor and estimate linear/nonlinear faults in the satellite attitude control system. Furthermore, a formation design is obtained in the presence of faults and disturbances from external sources. States and faults are combined to build a state-fault augmented vector. The observer utilized in this paper is an Unknown Input Observer (UIO) to decouple disturbances from fault and state estimations. We determine gain matrices using an H∞ approach to solve Linear Matrix Inequalities (LMIs). A numerical example is represented by three clusters of small satellites.