Open Collections

Abstract

In traditional fog-radio access networks (F-RANs), deploying a massive number of enhanced remote radio heads (eRRHs) connected to the macro base stations (MBSs) through fronthaul links is often constrained by site limitations and high deployment costs. To address these challenges, we propose a novel framework leveraging smart helpers (SHs), which operate independently of fronthaul links by smartly caching and delivering popular content based on within-coverage eRRHs' communications. This framework allows flexible deployment in diverse scenarios, such as dense urban areas and temporary public events, enhancing user quality of service (QoS) and reducing network load. We further extend this concept by integrating unmanned aerial vehicle-based smart helpers (UAVSHs) into F-RANs. To address UAV limitations, such as restricted battery life and coverage, we incorporate reconfigurable intelligent surfaces (RISs) to optimize the service area. For both ground-based and UAV-assisted SH systems, we formulate optimization problems to minimize average content delivery delays while optimizing cache resources, user scheduling, and RIS configurations. Multi-agent reinforcement learning (MARL) algorithms are developed to achieve these goals. Simulation results demonstrate substantial improvements in delivery delays, fronthaul load, and cache hit rates over traditional F-RAN architectures. In addition, we explore the optimal placement of SHs to maximize cache efficiency and minimize delivery latency. By modelling SH cache hit rates as functions of outage probability and user density distribution, we employ the radial basis functions (RBFs) method to estimate user density and optimize SH placement using the particle swarm optimization (PSO) algorithm. Our numerical analysis confirms the efficacy of the proposed methods, highlighting their potential to improve F-RAN performance by significantly enhancing cache hit rates, alleviating fronthaul loads, and reducing content delivery delays.