Research On The Optimization Of Digital Signal Processing Algorithms For Coherent Optical Communication Systems

Due to the advantages of large capability and long distance,the coherent optical communication holds an essential position in this high-speed information transmission age.However,since the coherent optical communication system has many defects,such as the dispersive in fiber,nonlinear effects,sampling clock error in ADC,frequency offset,phase noise,and so on,the received signal is damaged seriously.In order to solve these problems,the digital signal processing(DSP)technology is used in the coherent optical communication system to compensate these defects.Nevertheless,due to the increasing of transmission rate in coherent optical communication,the actual implementation of DSP technology faces a big challenge.In the practical optical fiber communication systems,the DSP algorithms shoud achieve real-time processing.Therefore,in order to reduce the requirement of the realization in hardware,the complexities of DSP algorithms is required to be reduced.To reduce the complexities of DSP algorithms,this dissertation concentrates on the research of DSP algorithms in coherent optical communication systems and the major research achievements and contributions are summarized as follows:(1)An M-order absolute operation is proposed to remove the modulated phase information.The proposed algorithm is applied to the mainstream frequency offset estimations(FOEs)based on fast Fourier transform(FFT)and phase difference to eliminate the interference of modulated phase information on FOEs.By researching,it is found that the proposed algorithm not only has high estimation accuracy but also reduces the complexity of FOEs.(2)A multiplier-free carrier phase estimation(CPE)algorithm is proposed.As one of the most common CPE algorithms,the Viterbi&Viterbi(V&V)algorithm is improved in this paper and a multiplier-free CPE algorithm is proposed.The proposed algorithm is investigated in the QPSK,8QAM,and 16QAM systems for CPE.The results show that the proposed algorithm has higher tolerance to noise than the V&V algorithm and requires no multiplier for which the complexity is reduced.(3)A FFT-based frequency offset estimation(FOE)algorithm with multi-steps interpolation is proposed.The proposed algorithm can effectively reduce the estimation error caused by the picket-fence effect of FFT-based FOE.The proposed algorithm has lower complexity than the least-squares FOE algorithm.Under the condition of same estimation accuracy,the complexity required by the proposed algorithm is just 7.26%of that required by the least-squares FOE algorithm.In addition,the proposed algorithm can reasonably select the number of interpolation steps according to the requirements of the tolerance to SNR and complexity,which improves its flexibility.(4)A combined FOE algorithm based on argument-FFT is proposed.Compared with the traditional FOE algorithm based on argument-FFT,the proposed algorithm requires only once operation of FFT,which effectively reduces the computation complexity.When the block length of data is greater than 213,the complexity of the proposed algorithm is half of that of the traditional FOE algorithm based on argument-FFT.(5)A CPE algorithm based on error function calculation is proposed.Through analyzing the reasons of the high complexcity of blind phase search(BPS)algorithm,a CPE algorithm based on error function calculation is proposed.In this proposed algorithm,a new error function is definded and the phase estimation value can be obtained by calculating the error functions of just four test angles.The proposed algorithm can also be combined with the improved V&V algorithm.This combined algorithm can approximately reach the same estimation accuracy as the BPS algorithm with 32 test angles,but the complexity of the combined algorithm is just 20.6%of that required by the BPS algorithm with 32 test angles.This dissertation focuses on the optimizations of FOE and CPE algorithms.These algorithms proposed in this dissertation can effectively reduce the complexcities of the FOEs and CPEs,which save the space and power of the realization in hardware and promote the development of real-time DSP technology in coherent optical communication systems.