基于异质驾驶行为的交通流微观建模与环境影响研究

发布时间：2018-11-03 17:07

【摘要】：2015年12月,《中共中央关于制定国民经济和社会发展第十三个五年规划的建议》中提出了"一带一路"建设,再次聚焦交通问题,同时也表明经济发展离不开交通的发展。为对交通流进行深入研究,学术界相继提出了宏观连续模型、跟驰模型和元胞自动机模型等交通流模型。在这些模型中,跟驰模型将车辆处理成相互作用的分散粒子,使其便于在计算机中进行模拟仿真,故应用广泛。在实际交通流中驾驶员并不是同质的,驾驶员异质性是普遍存在的。在心理学体系中,人与人之间存在着较大的差异性,例如性别、年龄、社会经济水平、受教育程度、健康状况等个体差异。在上述的变量中,有些与交通流建模相关,其直接地影响被驾驶车辆在给定操作环境和时间下的路径和速度。虽然驾驶员异质性的研究有很多,但是并没有先从驾驶员的异质性出发,多数先是统一处理数据,再进行异质性分析。鉴此,本文以大量驾驶员的轨迹数据为基础,拟对驾驶员行为进行深入研究,并提出科学的驾驶员分类方法,进而建立参数随机分布的车辆跟驰模型,并对不同类型车辆的环境影响因素进行分析。本文主要工作如下:(1)本文详细介绍了驾驶员异质性、车辆跟驰模型以及车辆油耗和污染物排放的相关研究内容。基于大量机动车数据(NGSIM数据),采用聚类分析的方法,对驾驶员进行分类。基于分类结果的特征,将驾驶员命名为"鲁莽—迟钝"型、"鲁莽—敏感"型、"谨慎—迟钝"型和"谨慎—敏感"型四类。(2)本文以GHR跟驰模型为基础,对车辆跟驰模型的参数分布特征进行深入研究。通过多元线性回归分析,得到了不同类驾驶员下GHR模型中各个参数的分布特征,并结合驾驶员类型进行了分析。(3)本文介绍并应用了 VT-Micro模型,然后根据油耗排放的回归系数计算了不同类型驾驶员的油耗分布、二氧化碳排放分布以及氮氧化物排放分布。通过其分布特征发现具有"谨慎"属性的驾驶员无论在油耗还是污染物排放量方面普遍高于"鲁莽"类驾驶员。反观具有"鲁莽"属性的驾驶员,其在油耗及排放方面的表现较为稳定,波动性远小于"谨慎"类驾驶员。
[Abstract]:In December 2015, the "proposal of the CPC Central Committee on formulating the 13th Five-Year Plan for National Economic and Social Development" proposed the construction of "Belt and Road" to focus again on traffic problems, and also showed that economic development can not be separated from the development of transportation. In order to deeply study traffic flow, the academic circles have put forward traffic flow models such as macroscopic continuous model, car-following model and cellular automata model. In these models, the following model treats the vehicle as an interacted dispersed particle, which makes it easy to simulate and simulate in the computer, so it is widely used. Drivers are not homogeneous in actual traffic flow, and driver heterogeneity is common. In the psychological system, there are great differences between people, such as sex, age, social and economic level, education level, health status and other individual differences. Some of the above variables are related to traffic flow modeling which directly affects the path and speed of the driven vehicle under a given operating environment and time. Although there are many researches on driver heterogeneity, most of them deal with the data uniformly and then analyze the heterogeneity. In view of this, based on a large number of driver trajectory data, this paper intends to conduct a deep study on driver behavior, and puts forward a scientific driver classification method, and then establishes a vehicle car-following model with random distribution of parameters. The environmental impact factors of different types of vehicles are analyzed. The main work of this paper is as follows: (1) this paper introduces the driver heterogeneity, car-following model, vehicle fuel consumption and pollutant emission in detail. Based on a large number of vehicle data (NGSIM data), drivers are classified by cluster analysis. Based on the characteristics of classification results, the drivers are named as "reckless-dull" type, "reckless-sensitive" type, "cautious-dull" type and "cautious-sensitive" type. (2) this paper is based on the GHR model. The parameter distribution characteristics of car-following model are studied in detail. By multivariate linear regression analysis, the distribution characteristics of each parameter in GHR model under different drivers are obtained, and the driver types are analyzed. (3) the VT-Micro model is introduced and applied in this paper. Then, according to the regression coefficient of fuel consumption, the fuel consumption distribution, carbon dioxide emission distribution and nitrogen oxide emission distribution of different types of drivers were calculated. It is found that the drivers with "caution" attribute are generally higher than "reckless" drivers in terms of fuel consumption and pollutant emissions. On the other hand, the drivers with the property of "recklessness" are more stable in fuel consumption and emissions, and their volatility is much less than that of "cautious" drivers.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.2

【相似文献】