Valuation And Its Applications For The Riccati Equation-based Self-tuning Information Fusion State

Multisensor information fusion is also called multisensor data fusion, whichsynthesizes the local information from the same or different sensors, eliminates theredundancy and contradiction, uses the information's complementation, forms theidentical and perceptual description for the environment. So it raises the rapidity andcorrection of the intelligence system decision and the scientificalness of the program. Itis a multistage process, including to detect, assemble and estimate the multi-source data,thus raises the precision of the stated and characteristic estimation. It prevents thelimitation of one sensor, and can obtain more information to get the more accurate andreliable conclusion.For the multisensor linear time-invariant discrete stochastic systems with unknownnoise variance matrix, using the new method of solving the matrix equations forcorrelation function, the estimators of the noise variance matrix are obtained. And usingthe classic Kalman filtering method, based on the Riccati equation, under the optimalfusion rules weighted by matrices, diagonal matrices and scalars, three self-tuninginformation fusion Kalman estimators are presented respectively. The self-tuningdecoupled information fusion Wiener estimators weighted by scalars are also presentedfor state components. They can handle the fused filtering, smoothing and predictionproblems in an unified framework. Their convergence (asymptotic optimality) is proved,i.e. if the parameter estimation of unknown noise variance matrix is consistent, they willconverge to the optimal information estimators in a realization. Many simulationexamples for the target tracking systems show their effectiveness.