One-Shot Blind Frequency-Domain Nonlinear Equalization for Millimeter-Wave Fiber-Wireless IFoF Mobile Fronthaul System

Abstract

Nonlinear equalization (NE) plays a pivotal role in the analog intermediate-frequency over fiber (IFoF) mobile fronthaul (MFH) system demanding a high signal fidelity. This work proposes and experimentally validates a novel blind frequency-domain NE (FD-NE) approach for millimeter-wave (MMW) fiber-wireless (FiWi) IFoF MFH systems. This proposed blind FD-NE scheme relies on the FD two-block cascade Hammerstein NE structure and particular pairs of M-ary phase shift keying (MPSK) and M-ary amplitude shift keying (MASK) subcarriers. It features lower computational complexity, higher transmission capacity, and comparable equalization performances to the widely-employed training-assisted memory polynomial (MP) NE schemes. Moreover, it only requires a small fraction of subcarriers in a single received signal block to be modulation-format constrained, which renders one-shot blind learning of NE coefficients and thus achieves symbol-by-symbol channel tracking. The performance of the proposed NE method is experimentally evaluated in MMW FiWi IFoF MFH systems operating at 5G NR frequency range 2-1 (FR2-1) n258 band (24.25 GHz to 27.50 GHz) and FR2-2 (52.6 GHz to 71 GHz). The obtained experimental results show that our proposal can effectively improve the nonlinearity tolerances of coherent 25-GHz and non-coherent 60-GHz FiWi IFoF MFH systems transmitting multiple 64QAM/256QAM IF OFDM signals.