A Novel ANN Equalizer to Mitigate Nonlinear Interference in Analog-RoF Mobile Fronthaul

Abstract

We first analyze the nonlinearities in the analog radio-over-fiber (A-RoF)-based mobile fronthaul systems with a multi-user uplink load. The inter-band cross modulations (XMs) present in the common fiber channel cause severe inter-user interference. Then, we propose a novel complex-valued multi-level artificial neural network nonlinear equalizer (ANN-NLE) for single-carrier frequency division multiple access signal uplink transmissions. The inter-user interference can be mitigated by employing a multi-user ANN-NLE. Finally, we design an experimental test bed comprising a 60-GHz wireless link and an A-RoF link in order to verify the model predictions and thereby demonstrate the ability of the ANN-NLE to overcome the nonlinear intra-band and inter-band XMs in the case of simultaneous joint equalizations of all users (multi-point-to-single-point) as well as in the case of individual user equalizations (point-to-point). From the experimental and simulation results, we demonstrate that the ANN-NLE can substantially reduce the inter-user interference and communication bit error.