We propose a joint modulation format identification (MFI) and optical signal-to-noise ratio (OSNR) estimation method via analyzing the trajectory information of the received adjacent symbols. A uniform grid in the first quadrant of constellation diagram is designed for converting the trajectory information of the received adjacent symbols to the adjacent matrix of corresponding mapped graph structure. Then, we translate the OSNR-independent MFI scheme to an appropriate canonical vector searching optimization problem via improving on the classical cosine similarly algorithm, where the eigenvector associated with the largest eigenvalue of the adjacent matrix is served as discriminated-feature to identify five commonly-used modulation format types, i.e., polarization division multiplexing (PDM)-quadrature phase shift keying (QPSK), PDM=8 quadrature amplitude modulation (QAM), PDM-16QAM, PDM-32QAM, and PDM-64QAM. After exact modulation format type is obtained, we build the relationship between the second-largest eigenvalue of the adjacent matrix and the OSNR value via polynomial regression method, and the least-squares principle (LSP) is utilized to find the best fit line for obtaining the optimal polynomial coefficient. The proposed method is verified by 28 GBaud PDM-QPSK, PDM-8QAM, PDM-16QAM, PDM-32QAM, and PDM-64QAM optical signal numerical simulations, and the results show that could achieve excellent performance on various parameter analysis. Besides, a proof-of-concept experiment is also conducted. More specifically, the proposed method only need one training process and one matrix decomposition, which would provide a superior convenience for future optical performance monitoring.
Read full abstract