地铁站虚拟乘客行为模拟研究与实现

RESEARCH OF VIRTUAL PASSENGER BEHAVIOUR SIMULATION ON SUBWAY STATION AND ITS IMPLEMENTATION

地铁站台人群仿真可以增强地铁驾驶模拟器的视觉沉浸感,同时可以用于地铁站台公共场所的应急处理仿真中。但由于乘客行为复杂,影响因素众多,因此乘客智能行为的模拟一直都是地铁仿真系统中的难点。通过调研分析地铁站台上下车人员的行为特征,总结乘客在多种因素影响下的行为规律。通过加入权重因子改进RVO算法实现全局最优路径查询和动态障碍避碰,对行为选择进行量化分析,建立了一套完整、可拓展性较强的智能乘客行为系统。解决了目前地铁模拟器中乘客模型简单、行为单一、真实感较差的问题,提高机车模拟驾驶过程中的场景逼真度和司机沉浸感。

This crowd simulation on subway platform can enhance the subway driving simulator visual immersion, and can be used for subway station public places in the emergency response simulation. However, because of the complexity and amount of effect factors of passenger behaviors, the simulation is the difficulty of subway simulation system. In this paper we analyze the behavioral characteristics of passengers on subway station in detail, and get a conclusion of behavior rules under kinds of effect factors. We improve RVO algorithm through weight factor method to realize dynamic obstacles avoidance. The paper makes quantitative analysis of behavior choices and establishes a complete scheme of intelligent passenger behaviors system with extensibleness and expansibility. Solved the problem of subway simulator such as simple passenger models, scarcity of behaviors, poor realistic, improves fidelity of the scene and driver immersive feeling through driving the subway simulator.