Channel estimation in OFDM systems using neural network technology combined with a genetic algorithm

Abstract

Orthogonal frequency division multiplexing (OFDM) is a multicarrier modulation used in a variety of broadband systems, such as asymmetric digital subscriber lines, very-high-speed digital subscriber lines, digital video and audio broadcasting, and wide local area network standards. OFDM is a promising solution achieving high data rates in a mobile environment. In wireless communication systems, signal distortion and attenuation in transmission caused by multipath effects makes it necessary to obtain knowledge related to the channel impulse response using channel estimation to provide compensation. This study combined a back propagation neural network for the estimation of channel and compensation signals with a genetic algorithm to improve performance and the convergence rate. We compared bit error rates and the mean square error of the proposed neural network with that of the conventional neural network, the least square (LS) algorithm, and the minimum mean square error (MMSE) algorithm in existing OFDM channel environments. Our results demonstrate that the proposed algorithm outperforms the LS algorithm and is on par with the MMSE algorithm.