Integrating Independent Component Analysis with Hopfield Recurrent Neural Network to Estimate the Channel of MIMO-OFDM System

Abstract

Channel estimation is very important for MIMO-OFDM system, and because of the complexity of the electromagnetism circumstance in multipath communication environment and the sensitivity of OFDM modulation mode, which brings great difficulty for the channel estimation. In order to solve the difficulty of channel estimation problem in MIMO-OFDM system, this paper propose a semi-blind estimation method based on independent component analysis (ICA) algorithm combined with improved Hopfield recurrent neural network (HRNN) as a hybrid approach named ICA-HRNN. Then use the ICA-HRNN algorithm to estimate the channel information of MIMO-OFDM system. The simulation results show that, the ICA-HRNN algorithm can better adapt to the nonlinear characteristics of MIMO-OFDM system, and increase the estimation accuracy and the estimation speed, especially when the system has low SNR.