Channel Equalization for MIMO-FBMC Systems

In recent years, it is been a great challenge to have high data transmission rates with additional bandwidth at the same time in wireless communication systems. Multicarrier systems along with MIMO have provided good results in achieving high bandwidth and spectral efficiency. Recently Filter bank multicarrier systems (FBMC) have been implemented and provided better results in terms of spectral shaping of the subcarriers as compared to the traditional orthogonal frequency division multiplexing (OFDM) with cyclic prefix (CP). Consequently, the major observable difference between the two approaches is in frequency selectivity. In this research, we will present a modified neural network based algorithm (NN) which is based on mean-squared error (MSE) trained for MIMO-FBMC systems with QAM modulation (QAM). The algorithm presents a per-subchannel adaptive channel equalizer with low complexity. Practical channel information has been used in the simulations. Furthermore, the convergence characteristic curves of NN based equalizer per-subcarrier will be discussed and also how the proposed algorithm will be optimized and evaluated. Moreover, to elaborate equalization concepts more in detail the proposed equalizer will be implemented for classical OFDM-QAM system and results will be compared to the simulations performed for traditional least mean square (LMS) based per-subcarrier channel equalizer.