高密度行人流具有显著的各向异性特征，行人在各方向空间尺寸和运动属性的差异直接影响空间的动态需求和利用，使得人群表现出规律性的集群运动。现有行人流理论考量的各向异性因素单一，不能充分深入地反映密集人群流动规律；现有模型不同程度地对行人进行了各向同性简化，其真实性与应用有所局限。本项目采集不同流动状态下的行人微观行为与运动规律数据，探索各向异性的空间与运动表征方式，量化其与环境参数的作用关系，分析各向异性对宏观流动参数变化规律的影响，揭示基于行人微观运动表征的集群运动形成与演化机理。构建能够实现行人空间和运动各向异性的精细网格场域元胞自动机模型，综合行人微观行为与运动参数发展行人动力学模型定量验证方法，缓解模型精细程度与模拟效率之间的矛盾。本项目将进一步完善行人动力学理论，为行人动力学模型的设计、标定与验证提供依据，为指导行人设施与路径规划、预测预警密集人群事故提供理论依据和技术支撑。

Dense pedestrian flow has an obvious property of anisotropy. The discrepancy in space and movement among different directions affects dynamically the requirement and utilization of space, which results in the regular crowd motion. Unfortunately, the anisotropy considered in the existing pedestrian flow theory is limited, which does not reflect fully and deeply the crowd motion; and to certain degrees, the existing models assume an isotropic simplification, which leads to a limitation in the reality and application. Thus, this project will collect empirical data on the microscopic behaviors and motion of pedestrians in different flows, establish the spatial and movement representation methods for the pedestrian anisotropy, and quantify the relationship between the pedestrian anisotropy and the environmental parameters. Then, in this project, we will analyze the influence of anisotropy on the macroscopic flow parameters and the mechanism for the formation and evolution of the crowd motion based on the microscopic movement representation of anisotropy. This project will also establish a fine discrete floor field cellular automaton considering the pedestrian anisotropy and develop a quantitative verification method for pedestrian dynamic models, integrating the microscopic anisotropy behaviors and motion of pedestrian and balancing the precision and the efficiency. This project will further supply the theory of pedestrian flow and support the design, calibration, and verification of pedestrian dynamics models. Meanwhile, it will provide the necessary theory and technical support for the pedestrian facility and path planning, and crowd disasters prediction.