Deterministic Selection of Phase Sequences in Low Complexity SLM Scheme

TL;DR

This paper introduces a deterministic cyclic shifts selection method for low complexity SLM schemes, improving PAPR reduction in OFDM signals by optimizing correlation properties compared to random selection.

Contribution

The paper proposes a simple deterministic cyclic shifts selection method that guarantees optimal PAPR reduction when correlation variance is low, outperforming random selection in low FFT size scenarios.

Findings

Deterministic cyclic shifts improve PAPR reduction performance.

Cyclic shifts are more influential than rotations for PAPR performance.

Proposed method outperforms Class III SLM in simulations for small FFT sizes.

Abstract

Selected mapping (SLM) is a suitable scheme, which can solve the peak-to-average power ratio (PAPR) problem. Recently, many researchers have concentrated on reducing the computational complexity of the SLM schemes. One of the low complexity SLM schemes is the Class III SLM scheme which uses only one inverse fast fourier transform (IFFT) operation for generating one orthogonal frequency division multiplexing (OFDM) signal sequence. By selecting rotations and cyclic shifts randomly, it can generate $N^3$ alternative OFDM signal sequences, where $N$ is the FFT size. But this selection can not guarantee the optimal PAPR reduction performances. Therefore, in this paper, we propose a simple deterministic cyclic shifts selection method which is optimal in case of having low variance of correlation coefficient between two alternative OFDM signal sequences. And we show that cyclic shifts are highly dependent on the PAPR reduction performance than rotations. For small FFT size and the number of alternative signal sequences is close to $N/8$, simulation results show that the proposed scheme can achieve better PAPR reduction performance than the Class III SLM scheme.