Uncertain Observation Arma Signals Of The Multi-sensor Information Fusion Estimation

The problems of signal processing have widely application in control and com-munications. ARMA(autoregressive moving average) signal has widely application background in many questions, such as establishing system model,analysis of time series prediction of system and deconvolution and so on. However, in the fact there are many bad factors including hardware facilities becoming old, influencing of en-vironment or data losing in the communication net would lead to uncertain nature of observation, which can be described by a series of Bernoulli distributed random variable. so the general estimate algorithm lost the optimal nature, because of losing the data of observation. Researching estimate problem with uncertain ob-servation on ARM A signal has important significance not only in theory but so in practice. In this paper, we address the problem of distributed information fusion estimate for ARM A signal with uncertain observation. The main content include distributed information fusion estimate by the state space method, by the estimate of observational noise for the sensors with uncertain observation and by the the innovation analysis approach.The ARMA signal model with uncertain observation is translated into the state model. Using the projection theory, the local optimal estimators, including filter, predictor and smoother, for state and white noise, which only include filter and smoother, are derived in the linear minimum variance sense. Thus we can ob-tain local optimal estimators including filter, predictor and smoother for ARM A signals model with uncertain observation. Meanwhile, the cross-covariance matrix of different-step estimator errors between any two sensor subsystems is derived. And the optimal fusion estimator is derived for multi-sensor ARM A signals with uncertain observations based on distributed optimal weighted fusion estimation al-gorithms.Using the projection theory, we can obtain the estimator of the observation noise in ARM A signal subsystem with uncertain observation for the state space model. The filter and smoother of ARM A signals can be obtained by the uncertain observation equation. The cross-covariance matrix of different-step estimator errors between any two sensor subsystems is derived relide on the estimator of ARM A signal. And the optimal fusion estimator is derived for multi-sensor ARM A signals with uncertain observations based on distributed optimal weighted fusion estimation algorithms.Using the innovation analysis approach, a novel optimal filter is obtained for discrete-time ARM A signals with uncertain observation based on the white noise es-timate. And we provide the cross-covariance matrix of different-step estimator errors between any two sensor subsystems. And the optimal fusion estimator is derived for multi-sensor ARM A signals with uncertain observations based on distributed optimal weighted fusion estimation algorithms.