New Weighted Bit-Flipping Decoding Algorithm for LDPC Codes

Abstract

Abstract —Several Weighted BF (WBF) algorithms are investigated in this paper, and a novel modified reliability-ratio based WBF (MRWBF) decoding algorithm for LDPC codes is proposed. WBF algorithm considers the influence of parity information on the error metric. Based on WBF, the improved WBF (IWBF) algorithm further uses the feature of message reliability on the symbol decision. The reliability-ratio based WBF (RRWBF) can eliminate the defect of requiring off-line pre-processing in IWBF. Our proposed MRWBF further removes the step of calculating normalized weighted parameters in RRWBF and achieves more reduction in decoding complexity. The results of simulations show that the proposed algorithm is feasible, effective, and can achieve good decoding performance. Keywords-LDPC Codes; WBF (Weighted Bit-Flipping); Message Reliability formatting I. BF sketch is as follows: I NTRODUCTION The Bit-Flipping (BF) decoding algorithm, which was proposed by Gallager in 1962, and rediscovered by by Mackay and Neal in 1996 [1], is based on the hard decision of symbol. BF is simple and easy to be implemented, but has only limited decoding performance. The weighted BF decoding (WBF) algorithm uses both the check relationships and the reliability of receive message, therefore obtains a better decoding performance when compared with BF algorithm [2]. Furthermore, an improved weighted BF algorithm (IWBF) [3] was introduced by Zhang et al., which further thought over the effect of message reliability on the error metric, and thus gained better decoding performance. But the IWBF introduces a weighting factor when exploiting the credibility of symbols. IWBF’s performance depends significantly on the selection of weighting factor, and the factor has to be preset with off-line process. Considering this defect, the reliability-ratio based WBF (RRWBF) was proposed to eliminate that weighting factor [4]. This paper also proposes a modified reliability-ratio based WBF (MRWBF) algorithm, which doesn’t need to pre-compute any factor and removes the step of calculating normalized weighted parameters existed in RRWBF, therefore MRWBF can further reduce the computational burden. The rest of this paper is organized as follows. We first discuss WBF, IWBF and RRWBF algorithms in Section II, then our MRWBF algorithm is introduced in Section III. Simulations are carried out in section IV, followed by conclusions in section V. II. BF, WBF, IWBF