ABSTRACT
This chapter investigates the information fusion estimation problem for a class of multisensor multirate systems with observation uncertainties of multiplicative noises. Different sensors have different sampling periods that are integer multiples of the state update periods. A centralized fusion filter (CFF) in the linear minimum variance sense is designed by using the observations received from different sensors at each time. It involves a time-varying Kalman filter with time-varying observation dimensions. Individual sensors provide their local estimators including filters and predictors based on their own observations. Then, local estimators are transmitted to the fusion center for fusion estimate at each time. A distributed fusion filter (DFF) has better robustness and flexibility since it has a parallel structure, which is convenient for the detection and isolation of the faults. Compared with the centralized fusion estimator, the distributed fusion estimator has flexibility and robustness since it has the distributed parallel structure.
