A0296
Title: Signal estimation from quantized measurements with random matrices, time-correlated noises and miscellaneous attacks
Authors: Raquel Caballero-Aguila - Universidad de Jaen (Spain) [presenting]
Jun Hu - Harbin University of Science and Technology (China)
Josefa Linares-Perez - Universidad de Granada (Spain)
Abstract: The least-squares (LS) linear estimation problem of stochastic signals using quantized measurements with random parameter matrices and time-correlated additive noises is addressed. This scenario is examined under the influence of mixed network attacks, including random deception attacks and denial-of-service (DoS) attacks, using Bernoulli random variables to model the stochastic nature of these attacks. Through an innovation approach, LS centralized fusion filtering and smoothing algorithms are derived, using a covariance-based methodology and a prediction compensation strategy to mitigate the effects of DoS attacks. A simulation example is presented to illustrate the broad applicability of employing random parameter matrices, which effectively cover a wide variety of network-induced uncertainties and random failures, thus offering a more faithful representation of engineering realities. The numerical simulations further corroborate the effectiveness of the proposed estimation scheme and shed light on the impact of random attack probabilities on the estimation accuracy. In sum, the proposed algorithms contribute to the advancement of signal processing and network security research, particularly in scenarios involving quantized measurements, mixed uncertainties, time-correlated noises and miscellaneous network attacks.
