复杂建筑应急疏散的动态多路径规划及路径表达机制研究

Emergency evacuation of complex buildings has become a research hotspot in the field of emergency treatment. Current intelligent building monitoring and management systems are generally lack of effective decision support and navigation capabilities for the building evacuation. According to information such as passage structure within buildings, real-time environment changes and position of evacuees, how to provide quickly and effectively all evacuees with best escaping routes at the same time is one of the difficulties. The project will take full account of three-dimensional, dynamic, complex characteristics of the emergency evacuation environment of complex buildings, and introduce the 3DGIS theory and technology in the three-dimensional mode to study the key scientific issues among the complex building emergency evacuation procedure. .The main contents include: 1) a new multi-level 3D indoor network model integrated with geometric, topological and semantic information, and its automatic constructing algorithm; 2) dynamic multi-path planning algorithms considering time-varying information about the evacuation environment, which will achieve the global optimal results of evacuation route planning; 3) the three-dimensional graphical representation and natural language description mechanism for the evacuation path. The study on related theories and the methods of dynamic multi-path planning and path expression of emergency evacuation for complex buildings, will further expand the spatial analysis ability and application of 3DGIS, promote the development and practicality of the building emergency evacuation research, but also lay the foundation for the study on the city indoor and outdoor integrated network modeling and analysis.

复杂建筑应急疏散已成为应急处理领域的一个研究热点。目前的智能建筑监控和管理系统普遍缺乏有效的疏散决策支持和导航能力，如何依据事故建筑内的通行结构、实时环境变化以及逃生者位置等信息，快速有效地为所有逃生者同时提供最优逃生路线是其中的难点。本申请充分考虑复杂建筑应急疏散环境的三维、动态、复杂特点，引入3DGIS技术在三维模式下研究复杂建筑内应急疏散过程中的关键科学问题，主要研究内容包括：1）融合几何、拓扑及语义信息于一体的多层次三维室内网络模型及自动化模型构建算法；2）顾及疏散环境中时变信息的动态多路径规划，达到疏散路径规划结果的全局最优；3）三维路径的图形化表达及自然语言化描述机制。通过研究复杂建筑应急疏散的动态多路径规划及路径表达的相关理论和方法，将进一步拓展三维GIS的空间分析能力和应用范围，推动建筑应急疏散研究领域的发展和实用化，同时也为研究城市室内外一体化网络建模与分析奠定了基础。

针对目前智能建筑监控和管理系统普遍缺乏行之有效的疏散决策支持和导航能力，本课题顾及复杂建筑室内应急疏散环境的三维、动态、复杂等特点，综合引入3DGIS、空间数据库、智能规划、图论等相关理论和技术方法，采用理论分析与实验测试相结合的方式，系统研究了三维模式下复杂建筑室内应急疏散过程中的关键科学问题。主要研究工作包括：1）分析了室内应急疏散规划和导航的特点，提出了融合几何、拓扑及语义信息于一体的多层次三维室内导航模型，并设计实现了模型各个层次的自动化构建算法；2）针对顾及疏散环境中时变信息的动态多路径规划，探索了利用遗传算法、蚁群算法以及传统图论算法实现应急疏散路径规划结果全局最优，实验证明智能算法的反复迭代会极大降低路径规划效率，而在利用传统图论算法解决动态多路径规划问题，时间的引入将使问题复杂度极大提升；3）提出了集成语音、文字和图形等导航信息形式为一体的多模式导航方法，实现了三维导航路径的自然语言化描述机制；4）利用ArcGIS Engine、Matlab等平台开发了实验测试原型系统和算法测试平台。. 本课题研究成果为拓展三维GIS在应急疏散和室内导航领域的应用提供了借鉴，促进了复杂建筑应急疏散、室内定位导航等理论和技术的发展。