Abstract

Auxiliary management and control channel (AMCC) has been defined in ITU-T standards for efficient deployment of a wavelength division multiplexing passive optical network (WDM-PON). However, the introduction of AMCC brings additional interference to the PON signal, and it is critical to reduce its impact on transmission performance. Effective modeling of the interference is crucial for the interference cancellation process. In this article, we propose a new modeling method for interference signals by mathematically fitting the baseband modulated AMCC signals through a Gaussian mixture model (GMM). Furthermore, a joint demodulation receiver is developed, which performs equalization via hidden Markov model (HMM). The proposed receiver optimizes the digital signal process according to the new interference signal model and can demodulate PON and AMCC signals in parallel. It allows the elimination of an additional optical conversion or signal acquisition circuit for AMCC, thus simplifying the receiver architecture. We verified the proposed new method on an experimental platform i.e. a PAM4 modulated 50G-PON system. Experimental results show that, aided by the proposed joint demodulation receiver, the AMCC can be modulated with a maximum modulation index reaching 15% using baseband on off keying (OOK) modulation, and the maximum transmission rate can be increased to more than 20 Mbps at a suitable modulation index. Compared with the existing demodulation methods, the proposed method can bring up to 2 dB of gain in PON demodulation with a bit error rate (BER) of $\mathbf {3.8 \times 10^{-3}}$ . This extends the application of AMCC and allows effective wavelength management for multi-level modulation WDM-PONs.