Optimal Weighted Measurement Fusion State Estimators And Its Application

With the development of computer and communication technology, multisensor information fusion technology develops rapidly and becomes an active research focus in current information processing domain during the past late years. The objective of multisensor information fusion filtering are to find the fusion estimators for the state or signal under some optimal fusion rules, the accuracy of which is higher than that of each local estimator, based on the local measurements or the local estimates for the state or signal of each sensor.For the multisensor linear discrete time-invariant stochastic control systems with same and different measurement matrices and with correlated measurement noises, using the weighted least squares (WLS) method, based on Riccati equation, two weighted measurement fusion Kalman filtering algorithms are presented respectively in this paper. For the multisensor linear discrete time-invariant stochastic systems with same measurement matrices and with correlated measurement noises and correlated input and measurement noises, a weighted measurement fusion Kalman filtering algorithms is also presented. Based on the Klaman filter in the information filter form, it is proved that they are completely functionally equivalent to the centralized measurement fusion Kalman filtering algorithm, so that they have global optimality. The corresponding weighted measurement fusion steady-state Kalman filtering algorithms and the weighted measurement fusion state component decoupled Wiener estimate algorithm are also presented, which have the asymptotical global optimality, and their applications in ARMA signal measurement fusion Wiener filtering are given. The simulation examples for the tracking system show their effectiveness.