A modified PSO assisted blind modulation format identification scheme for elastic optical networks

Abstract

By using a modified particle swarm optimization (M-PSO) design, a novel and blind modulation format identification (MFI) scheme was proposed for elastic optical networks (EONs) in this study. Compared with the traditional PSO algorithm, the sight of each particle in our M-PSO algorithm was strictly restricted to perform search of local extremum points (LEPs) on density distribution histograms (DDHs). In the experiment, the distributions of LEPs on DDHs depended on signals’ modulation formats, and then were regarded as the key identification metric. The proposed MFI scheme did not need prior information of optical signal noise rate (OSNR) in the identification processes, and showed a natural resistance to frequency offset and line width. The feasibility of the proposed scheme was verified by 28 GBaud polarization division multiplexing (PDM)-BPSK/QPSK/8QAM/16QAM transmission system, where the simulation results showed that the lowest OSNRs required to achieve 100% identification rate were all lower than their corresponding 7% forward error correction (FEC) thresholds (BER=3.8e−3). Furthermore, the tolerance of the residual chromatic dispersion (CD), differential group delay (DGD) and nonlinear effect were also verified under the simulation conditions. The study results manifest that the proposed MFI scheme showed interesting robustness to both linear and nonlinear impairments. Moreover, we also demonstrated a relatively simpler calculation complexity under the proposed MFI scheme.