基于移动互联网浏览器实现地铁站逃生模拟训练是一种高效率的火灾逃生训练方式.然而,由于地铁站规模庞大且火灾情景复杂,在线逃生路径规划仿真平台模型因数据规模大,其在基于有限网络带宽传输以及渲染能力较弱的浏览器上运行时,速度将...基于移动互联网浏览器实现地铁站逃生模拟训练是一种高效率的火灾逃生训练方式.然而,由于地铁站规模庞大且火灾情景复杂,在线逃生路径规划仿真平台模型因数据规模大,其在基于有限网络带宽传输以及渲染能力较弱的浏览器上运行时,速度将非常缓慢甚至无法运行.为解决此问题,本文针对轻量级Web3D地铁火灾逃生路径在线规划平台实时在线关键技术进行了研究.首先,针对大规模地铁站BIM静态场景数据,通过语义和体素化成分检验的轻量化方法对其进行了轻量化处理.同时,针对大规模虚拟化身的在线渲染,基于数据拆分并灵活组合思想,通过对虚拟化身的几何体信息和虚拟化身的动画数据进行数据管理,实现了大规模虚拟化身在线渲染的轻量化处理,进而实现了轻量级人群可视化;其次,针对动态烟气数据,提出了基于烟气冗余消除和归一化的轻量化处理方法,并基于精灵纹理粒子系统构建了轻量级烟气场景,实现了轻量级烟气可视化;最后,基于上述一系列轻量化处理的Web3D地铁场景中的逃生路径规划问题研究,本文提出了基于虚拟足迹聚类的蚁群优化算法eAACO (evacuation based on adaptive ant colony optimization),该算法通过VR设备获取真实人群逃生路径,实现对路径数据筛选和聚类以形成骨干路径,并与蚁群算法(ACO,ant colony optimization)相结合,设计了逃生路径规划的最优方案.实验表明,上述关键技术的实现较好解决了大规模地铁站火灾逃生路径规划Web3D模拟平台的实时在线处理问题.展开更多
According to architectural structure of underground station and condition of passengers,the emergency evacuations are extremely important during emergency accidents caused in the stations.Based on the architectural fe...According to architectural structure of underground station and condition of passengers,the emergency evacuations are extremely important during emergency accidents caused in the stations.Based on the architectural features and situation of passengers in the station,Multi-Agent system is proposed to analyze passengers’behavior and elements which affect passengers’behavior.Base on the crowd mentality,the model of selective behavior of passengers’route was established to simulate passengers’actions during emergency accidents.It is feasible to find out the influence brought by crowd mentality affect the emergency evacuations,and the influence of remitting the situation from external factor through WebVR.At last,it provides references for emergency evacuation strategy.展开更多
文摘基于移动互联网浏览器实现地铁站逃生模拟训练是一种高效率的火灾逃生训练方式.然而,由于地铁站规模庞大且火灾情景复杂,在线逃生路径规划仿真平台模型因数据规模大,其在基于有限网络带宽传输以及渲染能力较弱的浏览器上运行时,速度将非常缓慢甚至无法运行.为解决此问题,本文针对轻量级Web3D地铁火灾逃生路径在线规划平台实时在线关键技术进行了研究.首先,针对大规模地铁站BIM静态场景数据,通过语义和体素化成分检验的轻量化方法对其进行了轻量化处理.同时,针对大规模虚拟化身的在线渲染,基于数据拆分并灵活组合思想,通过对虚拟化身的几何体信息和虚拟化身的动画数据进行数据管理,实现了大规模虚拟化身在线渲染的轻量化处理,进而实现了轻量级人群可视化;其次,针对动态烟气数据,提出了基于烟气冗余消除和归一化的轻量化处理方法,并基于精灵纹理粒子系统构建了轻量级烟气场景,实现了轻量级烟气可视化;最后,基于上述一系列轻量化处理的Web3D地铁场景中的逃生路径规划问题研究,本文提出了基于虚拟足迹聚类的蚁群优化算法eAACO (evacuation based on adaptive ant colony optimization),该算法通过VR设备获取真实人群逃生路径,实现对路径数据筛选和聚类以形成骨干路径,并与蚁群算法(ACO,ant colony optimization)相结合,设计了逃生路径规划的最优方案.实验表明,上述关键技术的实现较好解决了大规模地铁站火灾逃生路径规划Web3D模拟平台的实时在线处理问题.
文摘According to architectural structure of underground station and condition of passengers,the emergency evacuations are extremely important during emergency accidents caused in the stations.Based on the architectural features and situation of passengers in the station,Multi-Agent system is proposed to analyze passengers’behavior and elements which affect passengers’behavior.Base on the crowd mentality,the model of selective behavior of passengers’route was established to simulate passengers’actions during emergency accidents.It is feasible to find out the influence brought by crowd mentality affect the emergency evacuations,and the influence of remitting the situation from external factor through WebVR.At last,it provides references for emergency evacuation strategy.
