Fusion Kalman Filter With The Wiener Filter, Based On The Observation Of The Modern Time Series Analysis Methods

The objectives of multisensor information fusion filtering are based on the measurement information or the local estimate information for the system state or signal, provided by each sensor, under a certain optimal fusion rule, the fused estimation for system state or signal is obtained, whose accuracy is higher than that of each local filter.Using the modern time series analysis method, based on the autoregressive moving average (ARMA) innovation model, under the weighted least squares(WLS)method optimal information fusion criterion, two weighted measurement fusion Kalman filtering algorithms are presented for the multisensor linear discrete time-invariant stochastic systems with correlated measurement noises; for the multisensor linear discrete time-invariant stochastic systems with correlated input and measurement noises and correlated measurement noises, one new weighted measurement fusion Kalman filtering algorithm is presented. It is proved that they are functionally equivalent to the centralized measurement fusion Kalman filtering algorithm, so that they have asymptotic global optimality. the corresponding weighted measurement fusion Wiener state filter and component decoupled Wiener state filter are presented .The optimal weighted measurement fusion Wiener filter and Wiener deconvolution filter of multisensor single channel ARMA signals are also presented .Compared with the centralized measurement fusion Kalman filtering algorithm, the weighted measurement fusion Kalman filtering algorithms not only give the globally optimal estimation, but also can obviously reduce the computational burden, so that they are suitable for real time applications. Many simulation examples show their effectiveness.