In this paper, a novel distributed cooperative estimation framework for a formation flight of satellites is proposed. This framework is developed based on the notion ofsub-observers.Within a group of sub-observers each one is estimating certain states that are conditioned on a given input, output, and state information. In order to guarantee the ultimate boundedness of the estimation errors, a sub-observer dependency (SOD) digraph is introduced that is assumed to be acyclic. The overall estimation process is modeled by aweighted sub-observer dependency estimation (WSODE) digraph.By selecting an optimal path in the WSODE digraph, a high-level supervisor can then select and configure a set of sub-observers to successfully estimate all the system states. In presence of unreliable information due to large disturbances, noise, and actuator faults certain sub-observers may become invalid. In this case, the supervisor reconfigures the set of sub-observers by selecting a new path in the WSODE digraph such that the impacts of these uncertainties are managed and confined to only the local estimates of states and faults. This will consequently prevent the propagation of uncertainties to the entire estimation process and the performance degradations to the entire formation flight of satellites. Simulations are conducted for a five satellite formation flight system in deep space and the comparative results with a centralized Kalman filter (CKF) technique are shown to confirm the validity and advantages of our developed analytical work.
展开▼
