Modulation format and baud-rate identification using asynchronous single channel sampling based on CNN

Abstract In this paper, an asynchronous single channel sampling (ASCS) based joint modulation format and baud rate identification (MF-BRI) method is proposed for M-QAM optical systems using Convolutional Neural Network (CNN), which significantly reduces the cost and complexity of monitoring system setup. Based on the three influencing factors of OSNR, CD and DGD, three sampling rates are analyzed to generate ASCS images to build the database. The simulation and experiment accuracy of MFI and MF-BRI is above 99.8%. We also discuss the differences between two-channel sampling (TCS) and single-channel sampling (SCS).