On Channel Estimation and PAPR Reduction of an OFDM Signal Using Partial Transmit Sequences

Abstract

To resolve the intrinsic high peak-to-average power ratio (PAPR) problem of orthogonal frequency division multiplexing (OFDM) systems, partial transmit sequences (PTS) scheme has been proposed, which partitions each OFDM symbol into a number of disjoint subblocks and then rotates their phases to form a low-PAPR signal for transmission. At receiver, in order to recover the original signal correctly, the index of the optimal candidate signal should be transmitted as a side information embedded in data symbols. However, demodulating data symbols need channel estimates which requires that pilot symbols should be inserted uniformly into data symbols in each OFDM symbol. And for the sake of channel estimation, the phases of the pilot symbols cannot be rotated along with that of the data symbols. Unfortunately, fixing pilot symbols gives rise to a large performance loss to PAPR performance when using PTS. In this paper, we find that since we have already know the original pilot symbols and all the possible rotation phases, the channel can be estimated in a blind way. Furthermore, rotating pilot symbols separately from data symbols can achieve a better PAPR performance than rotating them together. Simulation results demonstrate the feasibility of the proposed method.