The Research Of Nonlinearity Compensation Algorithms In Coherent Optical Communication

To satisfy the ever-increasing demand of broadband data transmission service, theoptical fiber communication systems are developed toward “ultra-high speed, ultra-largecapacity and ultra-long haul (3U)” transmission systems. Although the transmission distanceand capacity have been greatly improved after Erbium Doped Fiber Amplifier (EDFA) andWavelength Division Multiplex (WDM) technologies were invented, new challenges likeaccumulation of ASE noises, chromatic dispersion (CD) and nonlinear distortions havebecome the bottle neck of the next-generation fiber optic communication system. Coherentoptical transmission system has several merits such as high sensitivity, accurate wavelengthselectivity, compatibility with different advanced modulation formats and high spectralefficiency et al. Moreover, in combination with DSP technology, the various distortions canbe compensated effectively in such systems. Thus digital coherent detection has beenconsidered as the key technique for3U systems.Now the linear distortions like CD and Polarization Mode Dispersion (PMD) can beeffectively compensated with DSP algorithms, but due to the complexity of nonlinear effects,the nonlinear distortions compensation becomes the main constraints of3U systems. In thisthesis, digital signal processing (DSP) algorithms are intensively investigated to compensatefor nonlinear distortions impairments in coherent optical transmission systems. First, thedevelopment of the DSP based compensation technologies are comprehensively analyzedand summarized. Second, the mechanism and impact of the nonlinear effects in optical fibersare studied based the theoretical model. Their influences on3U transmission systems withadvanced modulation formats are also analyzed. Thirdly, the digital backward propagation(DBP) algorithm is investigated to compensate the intra-channel nonlinear distortions due toself phase modulation (SPM). The performance of DBP is systematically analyzed underdifferent sampling rates and residual CD values. The key parameters are optimized toimprove the performance, the system Q factor can be improved about0.5~2.1dB after theoptimization. Forth, to reduce the computational complexity, improved DBP algorithm, i.e.folded DBP and low pass filtered DBP, are developed for dispersion managed andun-managed systems. The impact of the deviation of the parameters from optical value isreduced. With equivalent performance the computational complexity can be reduced by a factor of ten. Finally the effectiveness of the DBP algorithms is verified using VPITransmissionMaker and Matlab for3U transmission systems with polarization multiplexedQPSK and16QAM formats.