Joint wireless resource allocation and bitrate adaptation for QoE improvement in IRS-aided RSMA-enabled IoMT streaming systems

Abstract

In the context of the increasing demand for Internet of Multimedia Things (IoMT) services, rate splitting multiple access (RSMA) and intelligent reflecting surface (IRS) technologies have been considered potential networking enablers to provide ultra-throughput wireless access. However, challenges arise due to the heterogeneity of IoMT devices and arbitrary network quality changes, resulting in unwanted service quality fluctuations and downgradation. This study addresses this problem by jointly optimizing wireless resource allocation and bitrate adaptation with Deep Reinforcement Learning (DRL)-based QoE management for IRS-aided RSMA-enabled IoMT streaming systems. We formulated the problem as a Markov decision process (MDP) and apply Proximal Policy Optimization (PPO) method to flexibly adjust IoMT bitrate, transmission beamforming, IRS phase shift, and RSMA parameters. As a result, our algorithm mitigates overestimation of client-side bandwidth, leading to smoother playback and reduced quality fluctuations. Simulations show that our approach outperforms baseline methods in terms of video resolution (up to 2.5 times) and achievable sum-rate (up to 50%), contributing to a superior streaming experience in IoMT systems.