面向元宇宙移动增强现实应用的异质内容 主动缓存与个性化交付机制  

Proactive caching and personalized delivery mechanism of heterogeneous content for MAR applications in the Metaverse

作　　者：徐思雅[1] 付琦梦 郭少勇[1] XU Siya;FU Qimeng;GUO Shaoyong(State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]北京邮电大学网络与交换技术全国重点实验室,北京100876

出　　处：《通信学报》2024年第10期55-68,共14页Journal on Communications

基　　金：国家自然科学基金资助项目(No.62201074);北京邮电大学基本科研业务费基金资助项目(No.2023ZCTH11)。

摘　　要：针对移动增强现实(MAR)应用异质数据内容传输和响应时延的问题,提出了面向元宇宙移动增强现实应用的异质内容主动缓存与个性化交付机制,首先综合考虑了用户特征和边缘节点服务能力,提出了用户行为和资源感知的边缘协作服务域构建方法;进而,基于协作服务域设计基于存储空间划分和用户偏好预测的异质内容预缓存机制;特别地,针对前景内容引入了个性化推荐机制,并设计了前景内容的差异化策略。仿真结果表明,所提机制在缓存命中率、前景内容和异质内容的平均响应时延方面均优于NCPCR策略、CCS-AGP策略和AIEC-RSC策略。Addressing the issues of heterogeneous data content transmission and response latency in mobile augmented reality(MAR)applications,a proactive caching and personalized delivery mechanism of heterogeneous content for MAR applications in the metaverse was proposed.Firstly,considering user characteristics and edge node service capabilities,a user behavior and resource aware edge collaboration service domain(ECSD)construction method was proposed.Then,based on the ECSD,heterogeneous content pre-caching mechanisms based on storage space division and user preference prediction were designed.Specially,by introducing the personalized recommendation mechanism of foreground content,a differentiated delivery strategy of foreground content was designed.Simulation results show that the proposed mechanism outperforms NCPCR,CCS-AGP,and AIEC-RSC strategy in terms of cache hit rate,average response latency for foreground content and heterogeneous content.

关 键 词：元宇宙 移动增强现实 异质内容 边缘缓存 内容推荐 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]