Lassaad SMIRANI
University of Carthage,Tunisia
Title: Artificial Neural Network Interpolation For SC-FDMA Channel Estimation
Biography
Biography: Lassaad SMIRANI
Abstract
This paper proposes an efficient interpolation method for SC-FDMA channel estimation. The proposed method of alternate pilots "Alternated Pilots Method APM", is based on three layers neural network model. APM interpolates channel estimated coefficients for data received symbols from a reduced numbers of transmitted pilot symbols. As we know LTE-A standard had proposed sending one pilot symbol on each frame of data which causes loss of throughput. Moreover, LTE-A assumes that the channel does not vary during frame transmission period. The APM uses only less than a half of pilot symbol used by standard method to achieve channel estimation. The APM increases bit rate and tracks the channel time variation. It allows migration from supervised channel estimation to semi-blind one.