We propose a novel low-complexity artificial neural network (ANN)-based nonlinear equalizer (NLE) for coherent optical orthogonal frequency-division multiplexing (CO-OFDM) and compare it with the recent inverse Volterra-series transfer function (IVSTF)-based NLE over up to 1000 km of uncompensated links. Demonstration of ANN-NLE at 80-Gb/s CO-OFDM using 16-quadrature amplitude modulation reveals a Q -factor improvement after 1000-km transmission of 3 and 1 dB with respect to the linear equalization and IVSTF-NLE, respectively.

In recent years considerable attention has been devoted to the merging of radio frequency and optical fiber technologies aiming to the distribution of millimeter-wave (mm-wave) signals. This effort has given birth to the field of Radio over Fiber (RoF) technologies and systems. This sort of systems have a great potential to support secure, cost-effective, and high-capacity vehicular/mobile/wireless access for the future provisioning of broadband, interactive, and multimedia wireless services. In this paper we present a comprehensive review of mm-wave frequency RoF systems. In our integral approach, we identified the most important figures of merit of an RoF system, which is divided into three main subsystems: Central Station (CS), Optical Distribution Network (ODN) and Base Station (BS). In each subsystem, the most promising technologies are classified: downlink transmission techniques at the CS, ODN architectures, and optical configurations of the BS. The impact of technology choice on the overall system performance is discussed, and the figures of merit are studied and used to assess the subsystem's performance. Finally, we suggest technological opportunities and future developments that should be attracting the attention of researchers and developers.

A reconfigurable nonlinear equalizer (RNLE) based on inverse Volterra series transfer function is proposed for dual-polarization (DP) and multiband coherent optical orthogonal frequency-division multiplexing (OFDM) signals. It is shown that the RNLE outperforms by 2 dB the linear equalization in a 260-Gb/s DP-OFDM system at 1500 km. The RNLE improves the tolerance to inter/intraband nonlinearities, being independent on polarization tributaries, modulation format, signal bit rate, subcarrier number, and distance.

Coherent optical orthogonal frequency division multiplexing (CO-OFDM) has attracted a lot of interest in optical fiber communications due to its simplified digital signal processing (DSP) units, high spectral-efficiency, flexibility, and tolerance to linear impairments. However, CO-OFDM’s high peak-to-average power ratio imposes high vulnerability to fiber-induced non-linearities. DSP-based machine learning has been considered as a promising approach for fiber non-linearity compensation without sacrificing computational complexity. In this paper, we review the existing machine learning approaches for CO-OFDM in a common framework and review the progress in this area with a focus on practical aspects and comparison with benchmark DSP solutions.

A novel versatile digital signal processing (DSP)-based equalizer using support vector machine regression (SVR) is proposed for 16-quadrature amplitude modulated (16-QAM) coherent optical orthogonal frequency-division multiplexing (CO-OFDM) and experimentally compared to traditional DSP-based deterministic fiber-induced nonlinearity equalizers (NLEs), namely the full-field digital back-propagation (DBP) and the inverse Volterra series transfer function-based NLE (V-NLE). For a 40 Gb/s 16-QAM CO-OFDM at 2000 km, SVR-NLE extends the optimum launched optical power (LOP) by 4 dB compared to V-NLE by means of reduction of fiber nonlinearity. In comparison to full-field DBP at a LOP of 6 dBm, SVR-NLE outperforms by $\sim 1\mathrm{dB}$ in $Q$-factor. In addition, SVR-NLE is the most computational efficient DSP-NLE.