Optical Impairment Compensation in Fiber Communication Systems Based on Artificial Intelligence: A Comprehensive Survey

The global demand for high-speed communication has increased dramatically over the past few years when data beginning to dominating of the traffic according to the Cisco Visual Networking Index (VNI). Data traffic is triple between 2014 and 2020, mainly, due to developing applications that consume bandwidth such as cloud services, HD video, high quality of real-time video transmission, virtual- augmented reality (VR-AR), online- games (video games), exchange of multimedia via smartphones and the more like. In fact, in 2020, more than a million minutes of multimedia (video) content is transiting the IP network every second according to the VNI; and the demands will exceed the capability of the current (core) internet backbone systems, in which optical communications are the main infrastructure. In this paper, the focus was on reviewing the mechanisms used for the most important and most effective techniques used to increase the capacity of optical transmission systems, namely Nonlinear Compensation (NLC) which work to reduce the nonlinear impairments that represent the main intrinsic challenges and the main capacity limitations facing the optical systems. The traditional NLC techniques were determined based on the approximate solution of the Nonlinear Schrodinger Equation (NLSE) through Digital Back Propagation (DBP), or Split- step Fourier Method (SSFM). however, their implementation demands excessive signal processing resources, and high-level accurate knowledge. A completely new approach that uses artificial intelligence (AI) algorithms to identify and solve these impairments has been studied in this paper. Traditional NLC techniques are reviewed in the first part to mitigation the nonlinearities and estimate the quality of transmission (QoT). Whereas in the second part, we review the uses of AI techniques that have been studied in applications related to monitoring performance, reduce nonlinearity, and quantify QoT. Finally, this paper presents a summary with a conclusion and outlook for development and challenges in optical fiber communication systems where AI is predictable to represent a hot major role in the near future.