The Constant Modulus Algorithm For PM-QPSK Coherent Optical Communication System

Polarization multiplexing coherent optical communication system has a high sensitive and a long length of transmission. Moreover, the high capacity of system can be obtained by the use of multi-level optical modulation formats and polarization multiplexing. Besides,compensation of linear transmission impairment can be realized in the electronic domain. The adaptive equalization algorithms of coherent optical communication have been studied and simulated systematically in our research. The details are described as follows:(1) A adaptive step-size constant modulus algorithm(CMA) method is introduced and applied to the dynamic equalization for quadrature phase shift keying(QPSK) coherent optical communication system. The conventional CMA equalization scheme are almost applied fixed step size, the choose of step-size always a trade between convergence speed and residual error. However, based the method in this paper, the adaptive step-size variable is adopted to replace the constant step-size factor in CMA equalization. Moreover, the key factor of proposed methods have been analysis and selected to make sure the algorithm convergence.Simulation results illustrate that, the proposed algorithm reduce about 600 iterations to convergent compared with conventional CMA, as 18% of total iterations. What’s more, as the BER thresholds is chosen in 10-3, the proposed scheme can save about 1.7d B in optical signal noise ratio(OSNR) penalty when compared with conventional CMA.(2) A novel high precision adaptive equalizer(FAT) is proposed which has lower computation complexity of time domain equalizer(TDE) and higher precision of frequency domain error(FDE). On the one hand, the computation complexity of time-domain CMA is high. On the other hands, the equalization performance of frequency-domain CMA is poor. In the proposed method, the first stage of equalizer adopt FDE CMA method to do the roughly equalization. After that, some undesirable equalized signal will be picked out by equalization error measurement. At last, a TDE fixed CMA method is proposed and used in time-domain equalizer, giving the more precise equalization to the signal which are picked out in the error measurement. Simulation results show that there were a 6% decrease for computation complexity by FAT when compared with TDE. Furthermore, as the BER thresholds is chosen in 10-3, the proposed scheme can save about 2d B and 2.2 d B in OSNR penalty when compared with TDE and FDE.