Centralized and distributed Kalman filtering in multi-coordinate systems with uncertainties

Algorithms for dynamic centralized and distributed algorithms are developed for multicoordinate systems in which where the communication networks are assumed to have uncertainties. In centralized Kalman filtering, the local measurements and the local sensor characteristics are transferred through the communication network to the central processor to generate optimal central estimate. In contrast, the local measurements, together with the previous central estimate transmitted from the communication network, are locally processed in distributed Kalman filtering. The statistics of local optimal estimates are sent to the central processor to yield the optimal global estimate. The effects of communication network uncertainties are minimized in both the local and central estimation