On time domain co-channel interference suppression for SC-FDMA transmission

In this work basic co-channel interference suppression techniques for SC-FDMA transmission are evaluated. As there seems to be only limited prior work on this topic, this work aims to provide a reference for more sophisticated suppression approaches still to be developed. Hereby, for the filtering the minimum mean-squared error (MMSE) criterion is considered. For the case of incomplete knowledge of the statistics of the interferers, a simple modified MMSE approach is evaluated, where these interferers are modeled as additive white Gaussian noise (AWGN). However, simulation results show that assuming co-channel interference as AWGN leads to a severe performance degradation. As an alternative an adaptive approach is presented, using the method of least squares (LS). Without any knowledge of the co-channel interferers the filter coefficients are adapted to current interference conditions exploiting a known reference signal. A refinement of the filter can be achieved by a feedback of already detected data in adaption, which further improves the performance. Although frequency domain equalization is common in SC-FDMA systems, the design of the LS filter is done in time domain, as for reduced filter length this cannot be done in frequency domain. Therefore, some general considerations which are important for time domain filtering in SC-FDMA systems are pointed out.