With Packet Loss Status And White Noise Estimation And Its Application Research

State estimation problem, widely used in military, research, control engineering, project production and other aspects, is a focal point in the feld of informationfusion. Noise estimation problem is the core technology of the oil seismic explo-ration. An important application background of that is seismic exploration signalprocessing, Along with the development of automation, intelligent technology, theresearch on the system estimation problem is gradually in-depth. And state esti-mation and white noise estimation problem attracted much attention at home andabroad. However, in the practical application, because of the bearing capacity of thenetwork and the restrictions of communication broadband, and all kinds of interfer-ence in the external environment, the packet dropouts of both sides from sensors toestimators and from controllers to actuators are taken into account. This not onlymakes the estimation on the original system performance was not accurate enough,reduced the accuracy of the estimation, but also add the difculty of the state andsignal estimation.This paper is concerned with the estimation problem, and derived state estima-tors and with noise estimators, for systems with multiple packet dropouts. Wherethe phenomena of multiple packet dropouts are described by a Bernoulli distributedrandom variable. There we only depend on the data arrival probability but notneed to need whether the sensor measurements at a particular time. Transform thesystems with multiple packet dropouts into state space model of the normal systemsresorting to state augmentation. According to the aforementioned literature, stateestimators and white noise estimators for systems with multiple packet dropoutscan be derived.In this paper, Resorting to white noise estimators, state and signal estimatorscan be realized in systems model. The ARMA (Autoregressive Moving Average)signal estimation question can be solved based on the state estimators and white noise estimators and corresponding error variance matrix can been given in thelinear minimum variance sense. In the end of this paper, we also design state andwhite noise estimators of descriptor system. Converting into two reduced ordersubsystems by non-singular transformations, the state and white estimators of thesubsystems are designed based on projection theory. Further, the estimators fororiginal descriptor system are obtained. Finally, we compare our flters with Kalmanfltering algorithm, in simulations, which not only show its efectiveness, but alsointerpret the widely application of the state and white noise estimations in ARMAsignal and descriptor system.