In today’s digital and interconnected world, the rapid growth of Virtual Reality (VR) and Augmented Reality (AR) has transformed online entertain ment, e-commerce, education, and training. VR/AR offers excellent oppor tunities for immersive experiences that engage users entirely, inducing a state of ’flow’ in which they lose themselves in the experience. However, the user’s sense of flow is highly dependent on their Quality of Experience (QoE), a term which encapsulates users’ subjective experience with not only content but also playout quality, achieved by maintaining an acceptable Quality of Service level. This work focuses on enhancing playout quality for streaming 360-degree videos. This is achieved using strategies to match the streaming video data rates to the available network resources, ensuring smooth video playout. Three broad approaches are considered: viewport prediction to identify the most important data to transmit; bitrate adapta tion to adjust content sent to networkconditions; anddynamicmanagement of network resources to support QoS for streaming video. The approaches examined in this thesis can be used alone or in combination to improve play out quality and enhance user experience with AR/VR systems that promote the creation of a sense of flow for users. Firstly, this thesis proposes the User Profile-based Viewport Prediction using Federated Learning (UVPFL) algorithm for 360-degree streaming which leverages machine learning to predict the focus of a user’s attention based on user interaction and head movement during the video playout. This serves to identify the most impor tant content to deliver, allowing for personalised content which reduces the overall video bandwidth requirements. Viewport prediction also forms the basis for a bitrate adaptation algorithm. Secondly, a technique based on Model Predictive Control (MPC) is proposed for rate adaptation of streaming 360-degree videos to match predicted network conditions. This technique offers improved QoE to the user by improving throughput, reducing latency, and maintaining smooth video playout. Finally, a network-centric Network Slice Management and Monitoring Solution (NSM2) is proposed. This so lution uses machine learning within a 5G framework to manage VR/AR traffic by dynamically assigning network slices based on real-time demand for user data to optimise service delivery. This approach allows the net work to support high-quality VR/AR services by ensuring seamless content delivery under varying network conditions.