Divneet Singh Kapoor1,*, Amit Kumar Kohli2
Intelligent Automation & Soft Computing, Vol.26, No.3, pp. 439-446, 2020, DOI:10.32604/iasc.2020.013920
Abstract This research paper represents an intelligent receiver based on the artificial-neuralnetworks (ANNs) for a 4x1 space-frequency-block-coded orthogonal-frequencydivision-multiplexing (SFBC-OFDM) system, working under slow time-varying
frequency-selective fading channels. The proposed equalizer directly recovers
transmitted symbols from the received signal, without the explicit requirement of
the channel estimation. The ANN based equalizer is modelled by using feedforward
as well as the recurrent neural-network (NN) architectures, and is trained using
error backpropagation algorithms. The major focus is on efficiency and efficacy of
three different strategies, namely the gradient-descent with momentum (GDM),
resilient-propagation (RProp), and Levenberg-Marquardt (LM) algorithms. The
recurrent More >
Login
Register
