Data Reduction Adaptive Digital Back Propagation Algorithm

With the application of various advanced technologies, the record of transmission reach in optical fiber communication has been continuously improved. In long-haul systems, in order to provide a higher power budget, the transmit power is increased, resulting in more severe nonlinear effects. However, the digital back propagation (DBP) algorithm, as a commonly used nonlinear impairment compensation method, requires accurate channel parameters when performing compensation, which is impossible in practical applications. To solve this problem, an adaptive DBP (ADBP) method based on data reduction is proposed in this paper, which is called data reduction ADBP (DR-ADBP). In DR-ADBP, instead of using all the received samples, a block with a certain length of them is selected and then applied to the adaptive algorithm to reduce the overall complexity. After the accurate parameters are searched by the adaptive algorithm, the DBP operation for all samples is performed then to compensate for the impairment. The proposal of the adaptive algorithm is under the guidance of the analysis of the nonlinear impairment and the adaptive cost function, which is more in line with the characteristics of the optical fiber channel. The proposed method is verified in a coherent optical communication system. The results show that under different initialization nonlinear scale factors, the convergence of nonlinear coefficients can be achieved by DR-ADBP with fewer iterations, and the required running time is much lower than that of the previous ADBP.