A novel DFT-based method for clipping noise suppression in OFDM systems

It is well known that clipping the OFDM signals in digital part of the transmitter is one of the simplest methods to reduce the peak to mean envelop power ratio. However, it suffers from additional clipping distortion, peak regrowth after digital to analog conversion, and out-of-band distortion. Recently, to combat the effect of in-band distortion and peak regrowth, it is proposed that before clipping, oversampling is performed by padding the modulating sequence with zeros. In this paper, we propose a robust DFT-based method (DBM) to reconstruct the clipped samples and mitigate the clipping distortion in the presence of channel noise at the expense of bandwidth expansion. We show through extensive simulations that by slightly increasing the bandwidth of the system, we can significantly improve the performance while limiting the maximum of the analog signal. Furthermore, we compare the performance of the DBM and the channel coding methods. It can be seen that for lower bandwidth expansions, the DBM outperforms the channel coding methods at moderate SNR values while for higher bandwidth expansions, the channel coding methods seem to be more efficient. Furthermore, we introduce a hybrid system which outperforms both the DBM and the channel coding methods at most SNR values.