Application Of Ldpc In100Gbps PDM-DQPSK Systems And Research Of Improved Decoding Algorithm For LDPC

In100Gbps and above high-speed optical communication systems, polarization mode dispersion and nonlinearity have become major factors limiting the enhancement of performance. In addition to the use of dispersion compensation modules and back propagation method of nonlinear compensation for the damage caused, forward error correction (FEC) technique to compensate various injuries in high-speed optical fiber communication can be used. LDPC in high-speed optical communication has become a hot research field of communication, and this article is based on the above.LDPC code has many advantages, it is suitable for long codes, and error correction performance of well-structured LDPC code is better than Turbo codes and close to Shannon limit. Sparse matrix makes encoding complexity low, and approximate triangular matrix achieves a linear time encoding complexity. Soft-decision decoding method is used, which makes the most of probability information in channel and leads to a lower BER.This paper focuses on the application of LDPC codes in100Gbps high-speed optical fiber communication. The main work and conclusions are as follows:(1) Several types of LDPC codes are achieved including finite geometry codes, SCG code, QC quasi-cyclic codes, PEG codes and%rotation.Construction and correction performance are researched and numerical simulation are conducted. The results showed that:SCG codes and QC codes can ensure no four ring. Finite geometry codes, QC codes and π rotation code are easily implemented in hardware, and the basic sub-matrices of π rotation code can be generated only using a replacement vector, and then the entire matrix H can be generated, so it has less storage space.(2) Analysis of coding time complexity of Gaussian elimination and approximation triangle method is conducted, and the latter is less than the former when g is small. The transmitting information equations of belief propagation decoding algorithm are derived; decoding algorithm for several other species were described.(3)100Gbps PDM-DQPSK simulation platform is set up with optisystem software, performance simulation with π-rotation LDPC codes with different rates are conducted in above systems. Simulation results show that the rate of0.5π-rotation LDPC codes compared to0.75for the same kind of code will bring greater net coding gain (NCG), which can bring down OSNR to14.3dB while BER is10-2.And when we use the chi-square statistical probability model, and the data is can be further reduced by0.1dB. Meanwhile, simulations of the tolerance for chromatic dispersion (CD), polarization mode dispersion (PMD) and nonlinear (NL) are conducted.(4) An improved decoding algorithm with characteristic bits is proposed. The characteristic bits are applied to above system of decoding algorithm. We infer the distribution of noise according to flag bits, which has an impact on information initialization formulas, further improve the performance of decoding algorithm. Simulation results show that:the number of iterations can be reduced by25%to50%with this method.