Efficient interference cancellation detector for asynchronous upstream optical code division multiple access-passive optical network with mixed Poisson-Gaussian noise

Optical code division multiple access (OCDMA) is a promising candidate for next generation passive optical networks (NG-PON). OCDMA-PON can potentially provide all customers with a Gb/s-class BW upstream with inherent flexibility. Unfortunately OCDMA suffers from multiple access interference (MAI) and various detection noises. In this study, the authors consider developing efficient interference cancellation detection under mixed Poisson-Gaussian noise. The latter is more realistic than considering Gaussian or Poisson noise alone. The authors first start by developing a maximum likelihood framework of the detection process under Gaussian, Poisson and Poisson-Gaussian noises, respectively. Then, the authors develop the proposed interference cancellation detector for the Poisson-Gaussian noise case after deriving the conventional expectation-maximisation (EM) detector for the Poisson noise case. Finally, the authors simulate the proposed detector and compare it to the projected parallel interference cancellation, EM and other detectors to validate its superiority and enhanced performance.