Abstract 5G mobile fronthaul (MFH) network requires data transmission with large bandwidth and ultra-low latency. Orthogonal frequency division multiplexing/offset quadrature amplitude modulation (OFDM/OQAM) passive optical network (PON) system is the 5G MFH solution with low spectrum efficiency and large bandwidth. The OFDM/OQAM-PON transmission performance will be undermined by intrinsic imaginary interference (IMI) induced by chromatic dispersion (CD) and polarization mode dispersion (PMD). In this paper, we proposed an artificial neural network (ANN) based channel estimation (CE) algorithm, which can effectively reduce IMI by estimation of channel transfer function (TF). Simulation results show that the proposed algorithm can optimize the system performance, compared with the conventional LS method, the proposed algorithm can improve bit error rate optimization capability by an order of magnitude.

中文翻译：

---

5G移动前传网络IM/DD-OFDM/OQAM-PON系统基于ANN的信道估计算法

摘要 5G移动前传（MFH）网络需要大带宽、超低时延的数据传输。正交频分复用/偏置正交调幅（OFDM/OQAM）无源光网络（PON）系统是频谱效率低、带宽大的5G MFH解决方案。OFDM/OQAM-PON 传输性能将被色散 (CD) 和偏振模色散 (PMD) 引起的固有虚数干扰 (IMI) 破坏。在本文中，我们提出了一种基于人工神经网络（ANN）的信道估计（CE）算法，该算法可以通过估计信道传递函数（TF）来有效降低 IMI。仿真结果表明，与传统的 LS 方法相比，该算法能够优化系统性能，