CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Optimal linear filter from fading measurements with time-correlated additive noises and transmission random losses Authors:  Raquel Caballero-Aguila - Universidad de Jaen (Spain) [presenting]
Josefa Linares-Perez - Universidad de Granada (Spain)
Abstract: The signal estimation problem in multisensor systems has developed into an important research area, due to its significant relevance in numerous applied and theoretical fields. It is common knowledge that networked systems frequently have random flaws, which, if not appropriately addressed, are likely to harm the performance of the estimators. The assessment of mathematical models and the development of estimation algorithms accounting for these random phenomena have thus received a lot of research attention. The fading or degradation of measurements (caused, e.g., by physical equipment limitations or inaccurate measurement instruments) is one of the most common uncertainties in sensor networks. Under the assumption that the measurements are affected by the fading phenomena, as well as perturbed by time-correlated additive noises and exposed to random packet dropouts during transmission, a recursive least-squares linear filtering algorithm is designed using a covariance-based methodology and a compensation strategy based on measurement prediction. For this kind of system with packet losses, the measurement differencing method, typically used to deal with the measurement noise time-correlation, is not successful, since some sensor measurements are randomly lost and, consequently, cannot be processed. Therefore, an alternative approach, based on the innovation technique and the direct estimation of the measurement noises, is used.