A0418
Title: Covariance-based distributed fusion filter under random delays, packet dropout compensation and signal-noise correlation
Authors: Raquel Caballero-Aguila - Universidad de Jaen (Spain) [presenting]
Josefa Linares-Perez - Universidad de Granada (Spain)
Abstract: The estimation problem in sensor networks has received considerable attention due to its multiple fields of application, which demand the development of new mathematical models and algorithms to accommodate the effect of unavoidable network-induced uncertainties. Especially significant are random transmission delays and packet dropouts, which, if not properly handled, may compromise the performance of estimators. The aim is to address the distributed fusion estimation problem of discrete-time stochastic signals from multisensor measurements subject to uncertainties modelled by random parameter matrices and additive noises, which are cross-correlated at the same time and correlated with the signal at the same and subsequent time steps. The signal evolution model is assumed to be unknown and only the mean and covariance functions of the processes involved in the measurement equations are available. Random one-step delays and packet dropouts occur during data transmission to the local processors and the prediction compensation strategy is used to attenuate the effect of packet dropouts. Using an innovation approach, a covariance-based least-squares recursive algorithm for the local filtering estimators is designed. Then, these local estimators are combined at a fusion centre to generate the proposed distributed filter as the least-squares matrix-weighted linear combination of the local ones.
