Machine Learning-Based Linearization Schemes for Radio Over Fiber Systems

Abstract

This manuscript proposes a novel machine learning (ML)-based linearization scheme for radio-over-fiber (RoF) systems with external modulation. The proposed approach has the advantage of not requiring new training campaigns in case the Mach-Zehnder modulator (MZM) parameters are changed over time. Our innovative digital pre-distortion (DPD) was designed to favor enhanced remote areas (eRAC) scenarios, in which the non-linearities introduced by the MZM become more severe. It employs a multi-layer perceptron (MLP) artificial neural network (ANN) to model the RoF system and estimate its post-inverse response, which is then applied to the DPD block. We investigate the ML-based DPD performance in terms of adjacent channel leakage ratio (ACLR), normalized mean square error (NMSE), resultant signal root mean square error vector magnitude error (EVM RMS ), and complexity. Numerical results demonstrate that the intended DPD method is less complex and outperforms the orthogonal scalar feedback linearization (OSFL) scheme, which has been considered a state-of-the-art DPD technique. The proposal has the potential to effectively and efficiently compensate for the RoF nonlinear distortions, especially in a time-variant system, without needing new training campaigns.

Digital Predistortion Based Experimental Evaluation of Optimized Recurrent Neural Network for 5G Analog Radio Over Fiber Links

“…[ Furthermore, advancements in the field include an overthe-fiber-based DPD approach employing reinforcement learning, resulting in a remarkable 60% reduction in bit errors [46]. However, these approaches have primarily focused on single-channel scenarios, with limited evaluation against standardized third-generation partnership project (3GPP) metrics like error vector magnitude (EVM) and adjacent channel power ratio (ACPR) [47][48][49][50][51][52][53]. Earlier integration of Analog RoF into fiber wireless networks achieved success, but without linearization, limited to 64 quadrature amplitude modulation (QAM) at 25 Gbauds, showcasing the need for performance-enhancing techniques [54].…”

“…Machine learning (ML) approaches have gained increased attention due to the demand for enhanced linearization to achieve superior outcomes [33,35,38,40,[42][43][44]. This growing interest is underscored by recent research conducted by Pereira et al, who explored ML algorithms for linearizing electrically amplified Radio over Fiber (RoF) systems [45].…”

Wireless and Optical Convergent Access Technologies Toward 6G

“…Otherwise, it will be necessary re-training the linearization algorithm, which will generate considerable expenses, since the communication system must be turned off for re-training, leaving customers without coverage. To overcome this issue, a ML-based scheme able to generalize possible variations of the A-RoF response was proposed by our research group, enabling a non-recalibrated linearization technique [148].…”