突发事件下民航机场旅客人群的疏散疏导协同优化技术

As important parts of the civil aviation transport industry, airports serve a high volume of passengers with different personal properties every day. Once an emergency like terrorist attack happens in an airport, it will lead to serious consequences for the safety of the passengers. Emergency evacuations for the airport passengers are among the most difficult challenges faced by every state all over the world today. .Facing with the crowd management and evacuation of passengers in airport under emergencies, the project focuses on the impact of the heterogeneity of crowd and the event itself on the pedestrian and evacuation dynamics. We will study the basic laws of pedestrian movement dynamics under long distance running, walking as well as crawling by controlled experiments and field observations. We will analyze the mechanisms of avoiding static and dynamic obstacles, as well as taking cover for pedestrians during their movements. Base on the above findings, we will extract the basic rules and equations of pedestrian movement and build pedestrian evacuation models especially for emergencies. Besides, we will develop algorithms for automatic analysis of spatial-temporal characteristics of crowd density distribution, fast detection and positioning of crowd abnormal behavior based on surveillance videos and convolutional neural network. Finally, we will build a prototype system of crowd management and emergency aid decision-making by integrating all the above modules and finally forming the collaborative optimization technique to achieve effective and applicative crowd management methods like route planning and guidance quickly. The research outcomes will provide theoretical and technical basis for understanding evacuation dynamics, decreasing stampede risk and reducing the injuries and deaths in airport under emergencies. Thus, it provides scientific support for enhancing the emergency management level for civil aviation airports.

作为民航运输的重要组成部分，机场的客流量大、人员构成复杂、内部设施多样，一旦发生恐怖袭击等突发事件，将会对旅客的安全造成极大影响。如何借助科技手段对旅客人群进行高效的疏散疏导，是一个急需解决的重要问题。.本项目拟针对突发事件下民航机场旅客的安全管理和紧急疏散问题，充分考虑机场人群构成的异质特征，采用实验和理论建模相结合的方法，研究异质人群在紧急情况下的长距离奔跑、步行、匍匐前进等疏散行为特征及其运动动力学基本规律，分析行人在动态避障和寻找掩护方面的行为机制，构建紧急情况下的人群疏散模型，开发基于监控视频和卷积神经网络的人群密度自动分析和异常行为快速检测定位算法，建立机场人群管理和应急疏散辅助决策系统原型，形成人员疏散疏导方案的协同优化技术。项目成果可为深入理解突发事件下的人员疏散规律、提高民航机场人群安全管理水平提供理论依据和技术支撑，具有重要的科学研究价值和现实意义。

民航机场的客流量大、人员构成复杂、内部设施多样，突发事件下旅客人群的应急疏散是公共安全领域面临的重要问题之一。本项目揭示了民航典型负载人群、群组、学龄前儿童等异质人群在通道、瓶颈等场景中步行、奔跑时的疏散动力学规律，提出了可用于监测人群密度以及人员异常行为监测的深度学习算法，构建了多因素影响下民航机舱红外火灾探测预警模型，耦合控制实验和虚拟实验数据，研究了突发事件下的决策行为规律，建立了基于空间感知的人群运动决策模型，成果可为深入理解突发事件下的人员疏散规律、提高民航机场人群安全管理水平提供理论依据和技术支撑。. 项目总计发表学术论文33篇，完成博士学位论文4篇，硕士学位论文3篇。发表文章被SCI收录24篇，核心期刊2篇，会议论文7篇，授权发明专利1项、实用新型专利1项、外观设计专利1项，参加 8 次国际学术会议，组织了国内学术会议2次。