Robust, Low Complexity LLRs using the Generalised Likelihood Principle

In this work we consider the performance of turbo coded OFDM systems in the presence of co-channel interference (CCI) in reuse-1 OFDM systems such as 802.16d/e and LTE. Because of frequency selectivity and diversity mapping, CCI does not affect all the subcarriers of the desired user. Hence CCI behaves like Narrow Band Interference(NBI) with the number of affected subcarriers ranging from 10% to 30%. The NBI in frequency domain is similar to impulsive noise in the time domain. Three different ǫ-mixture models have been considered for modelling impulsive noise. In this work, we propose a low complexity LLR(Log-likelihood Ratio) for turbo codes which is robust under all the three model assumptions. The proposed LLR does not assume the knowledge NBI power, NBI pdf, fraction of subcarriers affected by NBI and the Signal to Noise Ratio(SNR). Simulation results indicate that the proposed method performs within 0.8-1dB from the optimal LLR which would have complete knowledge of the pdf parameters and it is computaionally less complex.