A Low-Complexity CRC-Based Decoding Algorithm for SLM-ML OFDM Systems

In this letter, a low-complexity selected mapping (SLM) algorithm without phase sequence index information, e.g., side information (SI), is proposed. Different from the SLM algorithm without SI using maximum likelihood (ML), the proposed algorithm uses cyclic redundancy check (CRC) to terminate the decoding calculation in advance. Meanwhile, a small amount of known information is adopeted to assist CRC and a simpler phase sequence set is applied to the SLM. The calculation complexity and bit error rate (BER) of typical algorithms are compared with different phase sequence groups under several different channel conditions. As a result, lower computation complexity is obtained with the same peak to average power ratio (PAPR) suppression effect and no BER increase.

I. Introduction

Orthogonal frequency division multiplexing (OFDM) has the advantages of anti-multipath and high spectral efficiency. It is used as the main modulation method in 3/4/5G wireless communication systems. The superimposed transmission of subcarriers can effectively improve the transmission rate. However, the superposition of multiple subcarriers will introduce high PAPR, which requires higher transmit power and larger linear range of power amplifiers to ensure stable data transmission [1], [2]. Several algorithms have been proposed to reduce PAPR [3]. At present, SLM is one of the best PAPR suppression techniques and it can be divided into SLM with side information (SI-SLM) and SLM without side information (NSI-SLM) according to the phase recovery method. Although the decoding of SI-SLM is simple, the BER performance of OFDM systems is severely restricted by SI [4], [5]. NSI-SLM avoids the transmission of SI, most of which use the ML decoding at the receiver [6]–[7], [8]. For NSI-SLM, [8] proposed an NSI-SLM algorithm with pilot phase sequence, which reduces the calculation complexity compared to the algorithm in [7]. However, this algorithm still needs to perform ML calculation on all phase sequences at the receiver. In this letter, an algorithm is proposed by using a small amount of known information to assist CRC to terminate the calculation procedure in advance. Moreover, a simpler phase sequence is applied in the proposed algorithm. Compared with the algorithm in [8], the algorithm proposed in this letter obtains computational gain without increasing the BER and PAPR.