High-reliability Encoding And Decoding Technique For Product Codes

Error correction code has been considered as one of the key technologies for next generation high-throughput,high-reliability information systems.The low-density parity-check(LDPC)code is widely used in these fields,as it not only has superior performance which is close to the Shannon limit,but also can be decoded using parallel iterative decoding algorithm to support high throughput.However,it is not sufficient for LDPC codes to satisfy the ultra low bit error rate(BER)requirement of nextgeneration ultra-high-speed communications due to the error floor phenomenon.Product codes have been introduced as an efficient mechanism for constructing long codes using two or more sets of shorter linear block codes,and can improve the error floor of LDPC code.At present,the product code scheme has the high complexity,large loss of code rate and small coding gain.Based on this,this paper proposes a highreliability product encoding and decoding technique.According to the analysis of the residual error characteristics of the LDPC code,it can be seen that the number of residual errors of the LDPC codeword at the same position is less,and the majority of the LDPC codewords are detectable.Based on this,this paper proposes an erasure-correction-enhanced iterative decoding for LDPCRS(Reed-Solomon)product code.The erasure location information is provided by marking the LDPC codewords whose check value is not zero,and the hybrid errorerasure-correction(HEEC)decoding is proposed for the outer decoding.And then,the performance of product code is improved using iteration which is completed by updating the priori information of LDPC decoding.The proposed scheme effectively reduces the error floor of LDPC code,and meantime reduces the code rate loss introduced by constructing the product code.Compared with the traditional error-onlycorrection decoding method for the outer code,the proposed hybrid scheme has obvious performance gain.Meanwhile,compared with the LDPC product code which uses BCH code as the outer code,the performance of the LDPC-RS product code with the same code rate is also significantly improved.Finally,the overall performance of the product code is improved through iteration.Furthermore,compared with the binary LDPC code,the non-binary LDPC(NBLDPC)code has the lower error floor and the lower decoding threshold,and the error bit distribution is more concentrated,i.e.an error symbol contains several error bits.So,it can achieve better error matching with RS code under the same Galois Field.Based on the above analysis,the non-binary LDPC code is used as the inner code to construct the product code.Compared with the binary LDPC-RS product code,the non-binary LDPC-RS product code has obvious performance gain.Finally,the novel iterative decoding method for non-binary product code is proposed which is completed by modifying the bit prior information of LDPC decoding,and it effectively improves the overall performance of the non-binary LDPC-RS product code.