Improved Soft-Assisted Iterative Bounded Distance Decoding for Product Codes

Abstract

Product codes are demonstrated as good candidates for the forward error-correction (FEC) codes in fiber-optic communication systems. As a kind of hard decoding, iterative bounded distance decoding (iBDD) is widely adopted by the practical product decoders since it leads to low complexity and power consumption while achieving high net coding gain (NCG). By exploiting the channel reliabilities, soft-assisted iBDD (SA-iBDD) can avoid flipping some bits that are miscorrected and thus it has better decoding performance. In this paper, we propose an improved SA-iBDD for product codes. Based on the hard decoding results and the channel reliabilities, a voting strategy is introduced to judge whether a component codeword should be corrected or not. Compared with the conventional SA-iBDD, the proposed one improves the decoding performance at the cost of negligible complexity increase and without any additional memory requirement.