A Novel Channel Estimation Scheme Combining Adaptive Supervised and Unsupervised Algorithms

Abstract

[Abstract]: Channel estimation is a fundamental operation in digital communication systems that could be done using supervised or unsupervised (blind) strategies. Supervised techniques estimate the channel parameters using training symbols (pilots) included in the data frame, while unsupervised methods acquire information directly from the received signals. In this paper, we compare both strategies and propose a novel hybrid channel estimation method which provides performance similar to that achieved with the supervised algorithm but using a significantly more reduced number of pilot symbols. Such a hybrid scheme uses the information obtained at the receiver during the frame synchronization to decide the algorithm to be used in the data retrieval.