Title: A filtering algorithm for systems with random transmission delays modeled by multi-state Markov chains
Authors: Maria Jesus Garcia-Ligero - Universidad de Granada (Spain) [presenting]
Aurora Hermoso-Carazo - Universidad de Granada (Spain)
Josefa Linares-Perez - Universidad de Granada (Spain)
Abstract: An unavoidable problem in communication networks is the existence of delays in the arrival measurements. The delay may be deterministic or random, although in the most practical cases, such as mobile communications, exploration seismology, between others, the delay is random, being modeled by a stochastic process. Traditionally, the randomly delayed measurements have been modeled by independent random variables. However, in real communication systems, current time delays are usually correlated with the previous ones; a reasonable way to model the dependence on the delays is to consider them as homogeneous Markov chains. In this context, signal estimation algorithms have been derived considering that the measurements can be delayed by one, two or more sampling times. We generalize this situation to the case of measurements that can be delayed by one, two or more sampling times. Specifically, the least-squares linear estimation of signals is addressed considering that the delays are modeled by homogeneous multi-state Markov chains. The linear filter of the signal is derived by using the information provided by the covariance functions of the process involved in the observations, as well as the probability distribution of Markov chain modeling the delays.