基于单元重用的海量DVE场景的渐进式对等网络传输关键技术

项目名称： 基于单元重用的海量DVE场景的渐进式对等网络传输关键技术

项目编号： No.61272270

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 张晨曦

作者单位： 同济大学

项目金额： 70万元

中文摘要： 目前，在互联网上构建的虚拟空间正变得越来越大越复杂，网络需要传载的数据量急剧膨胀，然而网络带宽与网速的增长却远远跟不上用户实时浏览虚拟世界的传输要求。本课题采用：1.利用单元重用机制将需要传输的虚拟世界进行轻量化预处理，避免重复传输现象；2.再对轻量化后的场景进行流式化编码，进一步减少网络负载；3.透过增量式SMLAOI拣选当前最小下载区域；4.基于Voronoi图来发现整个P2P覆盖网中的潜在场景下载源，并选择其中的最优场景下载源节点；5.实施带宽增强策略与QoS带宽分配策略；6.系统研究基于重用度的渐进式预下载策略与场景更新策略。本课题提出了全新的海量DVE场景网上实时漫游的因特网传输的解决方案，整条技术路线与各项关键技术均有创新性与亮点，主要体现在：1.传输场景时不仅考虑了区域的渐进性，还考虑到模型的渐进性与重用性；2.首次研究基于重用的P2P渐进式传输、预下载与场景更新算法。

中文关键词： 分布式虚拟环境；兴趣组网；轻量化；缓存管理；数据迁移

英文摘要： Nowadays, online virtual reality applications are becoming larger and larger in both scale and complexity, which requires higher loads for Internet transmission. However, both bandwidth and speed of Internet tansmission are restricting the online virtual reality applications from real-time transmission and rendering. In order to solve this problem, in this project we propose: (1) a lightweight preprocessing of massive 3D models in DVE by exploring and reusing those repetitive texture patterns or 3D elements to avoid duplicated transmission; (2) a GPU accelerated streaming of DVE scenes to further reduce networking transmission loads; (3) an incremental SMLAOI algorithm to cull the smallest downloading region each step during a viewpoint moving; (4) a new algorithm to locate all potential candidates to provide downloading services progressively based Voronoi Graph throughout entire P2P overlay network; (5) a new algorithm to select the best peers from all available downloading peers; (5) a new method of bandwidth amplification to optimize QoS bandwidth allocation for better transmission performance; (6) a systematical solution of progressive P2P transmission, pre-downloading and caching mechanism of DVE scenes based on reusing repetitive 3D elements. The proposed research opens up a new area with a novel m

英文关键词： Distributed Virtual Environment；Interest networking；LightWeight；Cache management；Data Migration