Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-time Extended Reality Video Transmission Optimization Based on Frame-priority Scheduling

Guangjin Pan, Shugong Xu, Shunqing Zhang, Xiaojing Chen, Yanzan Sun

TL;DR

This work addresses real-time XR video transmission over 5G by formulating a frame-priority-based radio resource scheduling problem aimed at maximizing transmission quality, defined as the weighted sum of non-transmitted frames. It introduces a multi-step DQN framework (MS-DQN) that reduces action space and uses a CNN-based dueling network to accommodate dynamic state and action sizes, with a detailed MDP reformulation covering state, actions, transitions, and rewards. The approach leverages frame importance (e.g., I-frames vs P-frames) and per-frame deadlines, achieving up to 80.2% improvement in transmission quality over strong baselines and demonstrating faster convergence. The results suggest that frame-aware, multi-step reinforcement learning can significantly enhance XR experiences in 5G networks, enabling scalable, low-latency, high-throughput XR transmission in edge-enabled architectures.

Abstract

Extended reality (XR) is one of the most important applications of 5G. For real-time XR video transmission in 5G networks, a low latency and high data rate are required. In this paper, we propose a resource allocation scheme based on frame-priority scheduling to meet these requirements. The optimization problem is modelled as a frame-priority-based radio resource scheduling problem to improve transmission quality. We propose a scheduling framework based on multi-step Deep Q-network (MS-DQN) and design a neural network model based on convolutional neural network (CNN). Simulation results show that the scheduling framework based on frame-priority and MS-DQN can improve transmission quality by 49.9%-80.2%.

Real-time Extended Reality Video Transmission Optimization Based on Frame-priority Scheduling

TL;DR

This work addresses real-time XR video transmission over 5G by formulating a frame-priority-based radio resource scheduling problem aimed at maximizing transmission quality, defined as the weighted sum of non-transmitted frames. It introduces a multi-step DQN framework (MS-DQN) that reduces action space and uses a CNN-based dueling network to accommodate dynamic state and action sizes, with a detailed MDP reformulation covering state, actions, transitions, and rewards. The approach leverages frame importance (e.g., I-frames vs P-frames) and per-frame deadlines, achieving up to 80.2% improvement in transmission quality over strong baselines and demonstrating faster convergence. The results suggest that frame-aware, multi-step reinforcement learning can significantly enhance XR experiences in 5G networks, enabling scalable, low-latency, high-throughput XR transmission in edge-enabled architectures.

Abstract

Extended reality (XR) is one of the most important applications of 5G. For real-time XR video transmission in 5G networks, a low latency and high data rate are required. In this paper, we propose a resource allocation scheme based on frame-priority scheduling to meet these requirements. The optimization problem is modelled as a frame-priority-based radio resource scheduling problem to improve transmission quality. We propose a scheduling framework based on multi-step Deep Q-network (MS-DQN) and design a neural network model based on convolutional neural network (CNN). Simulation results show that the scheduling framework based on frame-priority and MS-DQN can improve transmission quality by 49.9%-80.2%.
Paper Structure (19 sections, 19 equations, 7 figures)

This paper contains 19 sections, 19 equations, 7 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. XR Traffic Model
  4. Channel and Scheduling Models
  5. Problem Formulation
  6. frame-priority-based scheduling method
  7. MDP Reformulation of Problem $\mathbf{P}1$
  8. State
  9. Action
  10. Transition Probability
  11. Reward
  12. DQN Model Design and Training
  13. EXPERIMENTAL RESULTS
  14. DDPGcite:priority-DDPG
  15. PFPF
  16. ...and 4 more sections

Figures (7)

  • Figure 1: XR transmission system framework. Real-time XR video is rendered on edge XR servers and then transmitted to clients for playback. BS uses frame information to schedule packets and improve transmission quality.
  • Figure 2: This example demonstrates three types of state transitions. In slot $t$, the agent of MS-DQN makes two scheduling decisions and transmits two frames.
  • Figure 3: CNN-based Neural network design for MS-DQN. This model can be applied to state and policy spaces of different sizes.
  • Figure 4: Convergence performance of MS-DQN and DDPG algorithms.
  • Figure 5: Performance comparison of transmission quality between the proposed MS-DQN algorithm and baseline algorithms.
  • ...and 2 more figures