Modeling radio over fiber communication system using augumented real-valued time-delay neural network

Abstract

Radio over Fiber technology has emerged as a pivotal solution for meeting the escalating demands for high-bandwidth connectivity in modern communication networks, particularly in the context of burgeoning wireless services. By seamlessly integrating the advantages of optical fiber’s low-loss, high-bandwidth transmission capabilities with the flexibility and mobility of wireless communication, Radio over Fiber systems offer a compelling architecture for delivering a wide array of services, including mobile broadband, wireless sensor networks, and in-building wireless solutions. As the optical signal propagates through the fiber, it is susceptible to both linear and non-linear impairments that can degrade the system’s performance. Accurate modeling of these impairments is crucial for predicting and mitigating their impact on the Radio over Fiber link’s performance, ensuring reliable and high-quality signal delivery. In this paper augmented real-valued time-delay neural network has been used to model the non-linear behavior of radio over Fiber Communication system and its performance has been compared with traditional memory polynomial-based modeling techniques.