Lattice reduction aided MMSE-SIC detection for non-orthogonal frequency division multiplexing signals

A lattice reduction aided MMSE detection combined with successive interference cancellation (SIC) scheme is proposed in this paper to suppress the noise amplification in the commonly used MMSE detection of non-orthogonal frequency division multiplexing (NFDM) signals. By adopting Lenstra-Lenstra-Lovasz (LLL) algorithm to reduce condition number of the nearly singular modulation matrix in NFDM, the new scheme improves bit error rate (BER) performance of MMSE detection as expected. Another advantage of the new scheme is that lattice reduction adds very little complexity to on line decoding of NFDM signals in AWGN channels because the reduction part of it could be done beforehand. Simulations show that complexity of the new scheme is much lower than the recently proposed decoding algorithm of randomized rounding semidefinite relaxation (SDR) in NFDM system, while more competitively, the BER performance of the new decoding scheme outruns direct randomized rounding SDR in high signal to noise ratio (SNR) when the frequency span is small or the number of sub-carriers is large.