Modeling and Simulation for Pedestrian Behavior in Congested Circumstances

本课题的目的是，构建密度大于1人/m^2的拥挤环境下的行人行为模型，并通过模拟来考察行人流的各种特性。Osaragi等人构建了基于心理压力概念的行人行为模型。模型的未知参数估计使用观测数据从真实的行人行为。该模型能够描述不同行人特征引起的行人行为差异。山田等人构建了一个模拟系统，模拟了移动的用户在道路、铁路、车站和交通信号灯等城市空间中的行为。例如，它可以模拟每个移动的用户从他/她的家出发，沿着沿着行走，到达车站，并乘坐火车离开的情况。利用该系统，从呼叫丢失概率等性能特征的角度讨论了移动通信系统的基站定位策略，Takahashi等人建立了两个模型。(a)密度大于1人/m^2的行人行为模型。该模型采用了行人之间的优先级的概念，并设备的其他行人的速度合适的估计机制。仿真研究表明，该模型可以应用于拥挤的情况下，密度高达2人/m^2。(b)行人模型，包括成对和成组行人。在该模型中，如果不拥挤，成对的人沿着行走，但如果严重拥挤，则独立行走。模拟研究表明，成对和成组的存在显著降低了行人的平均速度，例如，当密度为0.8人/m^2时，行人的平均速度降低了20%。

This project aims to construct model(s) for pedestrian behavior in congested circumstances with density higher than 1 person /m^2, and to examine various characteristics of pedestrian flow by simulation.Osaragi and his group have constructed a model describing pedestrian behavior based on the concept of mental stress. Unknown parameters of the model are estimated using observed data from real pedestrian behavior. The resultant model can describe differences of pedestrian behavior arising from different walkers' characteristics. Through the results of simulations of the proposed model, they examine comfort and efficiency of pedestrian space.Yamada and her group have constructed a simulation system to model mobile user behavior in urban spaces covering roads, railways, stations and traffic signals. It can simulate, for example, situations where each mobile user starts from his/her home, walks along street, arrives at a station, and leaves on the train. Using the system, base station location policies of a moble communication system are discussed in terms of performance characteristics such as call loss probabilities.Takahashi and his group have constructed two models.(a) A pedestrian behavior model for the cases with density higher than 1 person/m^2. The model uses the concept of priority between pedestrians confronting with each other and devices a suitable estimation mechanism for the speeds of other pedestrians. Simulation studies reveal that the model can be applied to congested circumstances with density up to 2 persons/m^2.(b) A model for pedestrians including pairs and groups. In the model, pairs walk along with each other if it is not congested, but walk independently if it is heavily congested. Simulation studies reveal that the existence of pairs and groups significantly decreases the average speed of pedestrians, for example 20% less when the density is 0.8 pedestrians/m^2.