Channel estimation by inference on Gaussian Markov random fields

Riedl, Thomas JohannesChoi, Jun WonSinger, Andrew Carl2010-01-072010-01-072010-01-07200910.1109/ACSSC.2009.5470038https://infoscience.epfl.ch/handle/20.500.14299/45141In this paper, we discuss a novel method for channel estimation. The approach is based on the idea of modeling the complex channel gains by a Markov random field. This graphical model is used to capture the statistical dependencies between consecutive taps in time and delay. The sum-product algorithm is finally employed to infer a MAP channel estimate from given observations.Markov random fieldchannel estimationchannel modelinggaussian message passingChannel estimation by inference on Gaussian Markov random fieldstext::conference output::conference proceedings::conference paper