Ldpc Codes Fixed Bits Remove The Decoding Algorithm And High-capacity Optical Communication Applications

LDPC code, whose performance is approaching the Shannon limit, has been successfully applied in communication engineering. And the decoding algorithm, the maximum numbers of decoding iteration and the quantified method have become a major issue for further study.Today, network requires the bit error rate as low as to 10^-12 to 10^-15. In the paper, we describe the channel coding theorem, the Shannon limit and the coding gain at first, and introduce the optical system architecture. Then, we introduce the BF decoding algorithm, the WBF decoding algorithm, the sum product decoding algorithm and the minimum sum decoding algorithm.The paper proposes a new decoding algorithm named deleting fixed bits decoding algorithm. Although it only applies to the systematic code, it also has applications in the high-performance FEC of 100Gbps optical technologies.The principle of the decoding algorithm is: we set the special information bits are fixed and deleted when sending during the LDPC coding; insert the advance deleted bits in the receiving sequence to form a complete code to decode. And we keep the fixed bits have maximum likelihood value in the iterative decoding process. Some LDPC codes would be appearing error floor in the impact of trapping sets in high SNR. If we choose the trapping sets to delete, that we can break the trapping sets to reduce or eliminate the error floors.We can find the key trapping sets with simulating the LDPC code. And use our new decoding algorithm to decode can greatly improve the performance of LDPC decoding.We use the algorithm to simulate and analyze a (13299,11285) LDPC code, found that: (13299, 11285) LDPC codes has the performance that the NCG is nearly 9.1dB when the BER is 10^-11, redundancy is 18.1%, expecting that error floor is less than 10-12, and matches the OTU4 frame format in optical communications.