Low-complexity fast converged carrier-assisted phase retrieval with adaptive weighted factors

Carrier-assisted phase retrieval (CA-PR) schemes significantly reduce the hardware cost of the receiver by using a dispersive element and two photodetectors (PDs) to reconstruct the optical field signal. The application scenarios are restricted by its limited signal reconstruction accuracy and high computational complexity. In this paper, a carrier-assisted phase retrieval scheme with adaptive weighted factors (CA-PR-AWF) is proposed. The intensity information obtained iteratively and the known intensity information obtained by square law detection are weighted by introducing a weighted factor, and the magnitude of weighted factor value is adaptively adjusted according to the computed intensity error to improve the convergence speed and accuracy of the phase retrieval (PR) receiver in signal reconstruction. The approach reduces computational complexity and supports two carrier scenarios: central carrier-assisted phase retrieval with adaptive weighted factors (CCA-PR-AWF) and edge carrier-assisted phase retrieval with adaptive weighted factors (ECA-PR-AWF). Effectiveness of phase retrieval is verified using 30GBaud 16QAM signals over 100 km of SSMF. The simulation results show that, under the condition of -1 dB CSPR, the number of iterations (NOI) for CCA-PR-AWF and ECA-PR-AWF schemes are reduced by 33.33% and 28.57%, and the computational complexity are reduced by 30.34% and 32.83%, respectively, compared with the conventional CA-PR scheme. Further analysis of specific performance metrics demonstrates that the OSNR penalty is reduced by 1.4 dB and 0.7 dB for CCA-PR-AWF and ECA-PR-AWF schemes, respectively, while laser linewidth tolerances are improved to 3.3 MHz and 4 MHz, respectively.