本文提出了一种新颖的视频流动态优化框架，该框架利用网络数字孪生（Network Digital Twin, NDT）技术，应对无线网络条件波动带来的挑战。传统自适应流媒体方法往往难以应对网络带宽、延迟和丢包率的快速变化，导致用户体验不佳，表现为频繁缓冲和视频质量下降。我们提出的框架集成了一个精密的NDT，可对无线网络进行实时建模，并采用预测分析来预测近期的网络状态。该NDT利用机器学习技术，具体为随机森林（Random Forest）和神经网络（Neural Networks），在带宽可用性、延迟趋势和潜在丢包影响视频传输之前对其进行预测。基于这些预测，我们的自适应流媒体算法可动态调整视频比特率、分辨率和缓冲策略，从而确保不间断的高质量观看体验。实验验证表明，与传统流媒体方法相比，我们的方法在多种网络条件下可将缓冲时间减少多达50%，并提升分辨率，显著提高了体验质量（Quality of Experience, QoE）。本文强调了将数字孪生技术集成到多媒体传输中的潜力，为更具韧性、以用户为中心的视频流解决方案铺平了道路。
英文摘要：
This paper introduces a novel dynamic optimization framework for video streaming that leverages Network Digital Twin (NDT) technology to address the challenges posed by fluctuating wireless network conditions. Traditional adaptive streaming methods often struggle with rapid changes in network bandwidth, latency, and packet loss, leading to suboptimal user experiences characterized by frequent buffering and reduced video quality. Our proposed framework integrates a sophisticated NDT that models the wireless network in real-time and employs predictive analytics to forecast near-future network states. Utilizing machine learning techniques, specifically Random Forest and Neural Networks, the NDT predicts bandwidth availability, latency trends, and potential packet losses before they impact video transmission. Based on these predictions, our adaptive streaming algorithm dynamically adjusts video bitrates, resolution, and buffering strategies, thus ensuring an uninterrupted and high-quality viewing experience. Experimental validations demonstrate that our approach significantly enhances the Quality of Experience (QoE) by reducing buffering times by up to 50\% and improving resolution in varied network conditions compared to conventional streaming methods. This paper underscores the potential of integrating digital twin technology into multimedia transmission, paving the way for more resilient and user-centric video streaming solutions.
When Digital Twin Meets 6G: Concepts, Obstacles, and Research Prospects
School of Science and Technology, Hong Kong Metropolitan University（香港都会大学科技学院）；School of Intelligent Systems Science and Engineering, Jinan University（暨南大学智能系统科学与工程学院）；Nanjing University of Posts and Telecommunications（南京邮电大学）
