返回
类型 基础研究 预答辩日期 2017-10-27
开始(开题)日期 2013-09-26 论文结束日期 2017-07-26
地点 交通学院三楼会议室 论文选题来源 国家自然科学基金项目     论文字数 6.8 (万字)
题目 行人过街行为特性的双层动力学建模与仿真研究
主题词 行人过街交通,行人短时路径决策模型,离散选择模型,行人过街设施设计
摘要 在人类的历史中,步行一直扮演着交通出行方式中开端和终端的重要角色。在“以人为本”思想的指引下,步行交通愈来愈受到交通管理者、研究者和出行者的关注。作为全部交通参与者中最易受伤害的行人群体,如何保障其在出行过程中安全是所有步行交通研究的中心点和重点。在常规情况下,行人在交通环境中最为危险的情境发生于过街时,尽管有信号灯、人行横道线和其他保护性交通设施的存在,但是,行人始终面临着与其他交通方式的竞争。若可以提高行人过街的效率,不但可以优化交通设施权属分配,提高城市交通系统运行效率,同时,也可以相应增强行人过街安全性和舒适性,从根本上解决“中国式过马路”这一管理难题。基于此,论文以大量的、富有针对性的调查数据为基础,系统分析和研究行人过街行为和特性,分别对中观行人过街决策行为和微观行人过街运动行为建模,将两个模型有机结合,从而建立可靠的行人过街行为双层动力学模型,通过计算机仿真的手段,提出行人过街设施的设计方法和步骤,在理论和实践两个角度对提高行人过街交通的效率、舒适性和安全性都具有一定的指导意义。 本文从行人过街行为的现象入手,选用人工调查法与视频调查法相结合的调查方法,通过对南京市4处平面交叉口的过街行人进行调查,从宏观与微观两个角度分析了行人过街行为的时空变化特性。重点分析过街行人流的速度、密度和流量等宏观特性,深入探究过街行人在速度变化、方向变化、起始位置选择等方面的微观特性,最后总结和归纳行人过街过程中出现的自组织现象规则。研究结果显示,过街行人到达分布服从负二项分布,南京市行人的理想过街速度为1.34m/s。行人在过街时更倾向于保持常速状态,少有选择加速或减速状态。行人进入过街设施时更喜欢选择中间区域的位置,而终点位置选择与起点位置的直线偏差不会太大,同时,行人过街走行方向主要集中在80°至100°区间内,呈现出行人倾向于按直线行走的特点。并从群体的角度归纳了行人过街行为的六项规则。 在行人过街行为特性分析的基础上,根据过街情境下行人过街行为的理论框架,将行人过街行为区分为战略层、战术层和操作层三个层级。以OWA算子为基础建立了行人过街短时路径决策模型。将两两比较法与OWA算子相结合,用以解决行人过街短时路径规划中的多目标决策问题。定义行人过街路径的最小成本,确保用户在计算每条路径成本的过程中能够使用多种决策策略,并且在选择过程中允许替换路线。提出最小鲁棒性路径成本的路径规划算法用以解决不同决策方案下的最佳路径选择问题。同时,利用Matlab GUI平台进行了原型系统开发,可实现基本特性输入、相对权重设置以及决策策略选择,并生成相应的决策路径。 在行人过街行为特性分析和行人过街决策模型的基础上,进一步建立了行人过街运动模型。以离散选择模型为基础,对行人过街的微观运动行为进行了建模。重点针对行人个体过街行为结构中操作层的行为进行了建模,包括约束和无约束两种行为模式。约束行为模式是由于与其他个体的交互作用而产生的,用跟随模式和冲突模式表示。这些模型捕获了人群行为特征的自组织效应。设计了比较完整的模型验证程序,并通过计算机仿真和数理统计进行测试,用第二章数据分析结果进行验证和校准保证了模型在类似的情景下表现良好,并且在用于预测时发现模型特性具有鲁棒性。 以行人过街行为基本理论中战术层和操作层的交互为切入点,通过计算机仿真的手段,将对向行人流运行状态区域分为起步区、波动区以及交织区,并对状态判别区域范围进行量化。引入行走负效应概念,建立人行横道状态评价指标与分级体系,提出行人交织区的三级运行状态,并与HCM 2010手册中人行设施分级部分进行对比。同时,结合仿真数据,研究不同人行横道宽度与其通行能力的关系,进而提出人行横道宽度设计方法;对不同过街延误影响因素进行分析,综合过街时间,延误占比和局部密度水平等指标,研究负效应与延误的相关性,并以此指导行人过街最小绿灯时长设计。
英文题目 MODELING AND SIMULATION OF PEDESTRIAN CROSSING BEHAVIOR BASED ON BI-LEVEL DYNAMICS MODEL
英文主题词 pedestrian crossing; Pedestrian short-time path decision model; Discrete selection model; Pedestrian crossing facilities design
英文摘要 Walking has always played an important role as the beginning and end of tourisr in human history. Under the guidance of "people foremost", pedestrian traffic is becoming more and more concerned by traffic managers, researchers and travelers. As, The focal point and focus of all pedestrian traffic studies is how to ensure the safety of pestrian, which is the most vulnerable group in all traffic participants, travel. In normal circumstances, the most dangerous situation occurs when pedestrian cross the street. Although there are traffic lights, pedestrian crossing lines and other traffic facilities which could do the effect of protection, pedestrians still always face the competition from other modes of transport. Improving the efficiency of pedestrians crossing would not only optimize the ownership allocation of traffic facilities and improve the efficiency of the urban transportation system. Also, it could enhance the safety and comfort of pedestrian crossing accordingly. That would fundamentally solve the management problem "Chinese style of crossing the road". Based on this, the paper analysis and the characteristics of the pedestrian crossing behavior based on the sufficient and efficient survey data. Also, paper modeled the decision-making behavior in intermediate perspective and moving behavior in microscopic perspective of pedestrians crossing. With the combination of the two models, so as to establish a reliable pedestrian crossing behavior of bi-level dynamic model. Through computer simulation, the paper put forward to give design methods for pedestrian crossing facilities, which could improve the pedestrian crossing traffic efficiency, comfort and safety of all has a certain guiding significance in both theory and practice perspective. Based on the phenomenon of pedestrian behavior, investigation methods combined manual survey method and video survey method. Four intersections were surveyed in Nanjing and from the macro and the characteristics of the temporal and spatial variation for pedestrian crossing behavior were analysed in macro and micro level. Focus on the analysis of pedestrian crossing speed, density and flow of macro micro characteristics, the pedestrian behaviors in speed change, the change of direction, the starting position selection were explored. And the process of self-organization rules of pedestrian crossing street was summarized. The results show that the pedestrian arrival distribution follows two negative distributions, and the ideal pedestrian crossing speed is 1.34m/s in Nanjing. Pedestrians are more likely to remain at normal speed while crossing the street, and less likely to accelerate or decelerate. Pedestrian crossing facilities prefer to choose the middle position. And end position will stay on a straght line with start position. At the same time, pedestrian walking direction mainly concentrated in the 80°to 100°, showing people tend to travel according to the characteristic of linear walking. Finally, the six rules of pedestrian crossing were concluded from the point of view of community. According to the theoretical framework of pedestrian crossing behavior in pedestrian crossing, the pedestrian crossing behavior is divided into three levels: strategic layer, tactical layer and operation layer based on the analysis of pedestrian crossing behavior characteristics. The paper proposed the OWA operator to establish the shortest path decision model for pedestrian crossing. Combining two-two comparison method with the OWA operator to solve the multi-objective decision-making problem in short path planning of pedestrian crossing. Defines the minimum cost of pedestrian crossing paths to ensure that decision-makers can use multiple decision strategies in calculating each path cost, and allows alternative routes during the selection process. A path planning algorithm with minimum robustness path cost is proposed to solve the optimal path selection problem under different decision alternatives. At the same time, the prototype system is developed by Matlab GUI platform, which can complete the basic characteristic input, relative weight setting and decision strategy selection, and generate the corresponding decision path. Through pedestrian crossing behavior analysis and pedestrian crossing decision-making model, pedestrian pedestrian crossing model was further established. Based on discrete choice model, the micro motion behavior of pedestrian crossing was modeled. Particularly, the behavior layer in pedestrian pedestrian crossing structure was modeled, including two kinds of behavior modes: constraint and no constraint. The constraint behavior model is focus on the interaction between individuals, and was represented by following patterns and conflicting patterns. These models capture the self-organization effects of population behavior characteristics. Besides, a completed validation procedure was desgined and model was tested by computer simulation based on the data in second chapter. The analysis results showed a well performance of the model in similar situations and the characteristics of robustness in prediction. In the basic theory of pedestrian crossing behavior, the interaction between tactical layer and operation layer was taken as the breakthrough point. By means of computer simulation, the pedestrian area was divided into the starting zone, fluctuation area and weaving area. Besides, the range of state discrimination is quantized. The concept of walking negative effect was introduced, and the evaluation index and classification system of pedestrian crossing condition were established. The three running states of pedestrian weaving zone were put forward and compared with the classification of pedestrian facilities in HCM 2010 manual. Meanwhile, the relationship between different crosswalk width and capacity was studyed combined with simulation data. Then, a design method of different width of pedestrian crossing delay was put forward. Based on the relationship between negative correlation effect and delay, crossing time delay, and the proportion of local density levels were studyed and guiding the design methond for the minimum green time of crossing street pedestrian.
学术讨论
主办单位时间地点报告人报告主题
课题组 2017.04.20 东南大学交通学院会议室 于晨 共享单车
课题组 2013.3.24 东南大学交通学院会议室 张婧 复杂网络
课题组 2013.3.31 东南大学交通学院会议室 江航 元胞自动机在交通系统微观建模中的应用
课题组 2013.5.16 东南大学交通学院会议室 张婧 基于GIS-T的城市道路交通拥堵预警方法与疏导策略研究进展
课题组 2013.06.19 东南大学交通学院会议室 于晨 公共设施选址经典模型介绍
课题组 2014.11.20 东南大学交通学院会议室 张婧 基于蒙特卡罗仿真的交通拥堵趋势预测方法
课题组 2015.09.24 东南大学交通学院会议室 于晨 数据可视化
课题组 2016.10.27 东南大学交通学院会议室 于晨 交通流仿真模型与仿真软件介绍
     
学术会议
会议名称时间地点本人报告本人报告题目
GITSS2014 20140701-20140703 中国西安 Multi-stage Pedestrian delay model at signalized intersection
CICTP2016 20160705-20160709 中国上海 simulation of pedestrian flow at a crosswalk by multi-agent system with pre-avoid forces model
     
代表作
论文名称
Simulation of Pedestrian Flow at a Crosswalk by a Multi-Agent System with a Pre-Aviod Forces Model
 
答辩委员会组成信息
姓名职称导师类别工作单位是否主席备注
张卫华 正高 教授 博导 合肥工业大学
陈学武 正高 教授 博导 东南大学
郑长江 正高 教授 博导 河海大学
何杰 正高 教授 博导 东南大学
杨敏 正高 教授 博导 东南大学
      
答辩秘书信息
姓名职称工作单位备注
梁衡宏 其他 工程师 东南大学