应用智能公交和路网数据的城市公交站点出行计算模型与评价

发布时间：2018-05-03 20:22

本文选题：公交车 + GPS　；参考：《太原理工大学》2017年硕士论文

【摘要】：“智慧公交”是“智慧城市”的重要组成部分,是解决城市交通问题和方便居民出行的有效途径。智慧交通不仅可以诱导出行,还可以通过历史大数据的分析决策出行。公交客流量是深度挖掘交通出行大数据、研究乘客出行模式的基础。公交车到站时间更是出行者最为关心的交通信息之一。因此,以地理信息系统和数据分析为手段,展开对公交出行分析及挖掘工作,结合公交车数据结构,探讨乘客上下车站点推断和公交车到站时间预测方法,对城市交通问题的解决具有积极意义。本文在综合分析国内外对客流量和出行链研究方法的适用性、公交到站时刻模拟预测速度优缺点的基础上,结合数据源特点和人力财力,提出以单条出行链为研究对象,研究确定各站点吸引权,计算站点客流量;建立多元线性回归模型计算公交车历史平均车速,综合瞬时速度和到站距离,计算修正平均速度,预测公交车到站时间。基于深圳市AFC和GPS数据,利用时间匹配和密度聚类方法确定乘客上车站点;分析乘客出行行为以及规律,引入出行链单元公交节的概念。公交出行节连续时,依据乘坐人下次乘车的上车位置判断乘客下车站点;公交出行节断裂的乘客,结合乘客刷卡高频站点的频次和公交路线下游各站点吸引权重,判别出行节断裂时乘客下车位置坐标的可能性,并设计推断乘客下车站点算法。根据预测得到的乘客上下车站点信息,统计估算车内人数。利用K最邻近结点的方法对道路进行分段,建立多元回归速度模型估计各路段平均速度,以计算结果为历史数据依据,结合公交实时瞬时速度和距离到达站点的距离长度,预测公交的到站时刻。根据公交乘客下车站点推断算法,实例分析并预测结果,计算下游各站点的乘客可能的下车频次和分析乘客高频下车站点集,分析算法可行性,根据乘客下车预测点与真实下车站点之间的距离和各个预测点的权重判别评估预测的准确性,经过验证,表明方法是有效的。依据到站时间预测模型计算实际公交到站时间,通过与真实值对比评估,表明误差在合理范围内。利用路段平均速度的计算结果建立数据库,并对道路通畅性进行级别划分和实时可视化表达,其结论符合实际状态。
[Abstract]:"Smart bus" is an important part of "Smart City", which is an effective way to solve urban traffic problems and facilitate residents to travel. Intelligent transportation can not only induce travel, but also travel through historical big data's analysis and decision. Public transport passenger flow is the basis of deeply excavating traffic travel big data and studying passenger travel mode. Bus arrival time is one of the most concerned traffic information for passengers. Therefore, by means of GIS and data analysis, the analysis and mining of bus trip are carried out, and combined with the bus data structure, the methods of estimating the stop and the arrival time of the bus are discussed. It is of positive significance to solve the urban traffic problems. On the basis of synthetically analyzing the applicability of domestic and foreign research methods of passenger flow and trip chain, and the advantages and disadvantages of simulating and predicting the speed of bus arrival time, combined with the characteristics of data sources and human and financial resources, this paper puts forward a single trip chain as the research object. The research determines the attraction right of each station, calculates the passenger flow of the station, establishes the multivariate linear regression model to calculate the bus historical average speed, synthesizes the instantaneous speed and the distance to the station, calculates the revised average speed, and predicts the bus arrival time. Based on the data of AFC and GPS in Shenzhen City, the method of time matching and density clustering is used to determine the passenger boarding station, the travel behavior and regularity of passengers are analyzed, and the concept of public transport section of trip chain unit is introduced. When the bus travel section is continuous, the passenger gets off the bus station according to the passenger's next boarding position; the passengers whose bus trip node is broken combine the frequency of the high-frequency station and the attraction weight of the lower reaches of the bus route. The possibility of determining the coordinates of the passenger's alighting position when the trip node is broken and the algorithm of inferring the passenger's stopping station are designed. According to the forecast of the passenger station information, statistics estimate the number of people in the car. Using the method of nearest node of K to segment the road, a multivariate regression speed model is established to estimate the average speed of each section. Based on the historical data, the real-time instantaneous speed of public transportation and the distance length from the arrival station are combined. Predict the arrival time of the bus. According to the algorithm of bus passenger station inference, the example analysis and prediction result, the calculation of the possible frequency of passengers getting off the lower reaches and the analysis of passenger high frequency station set, the feasibility of the algorithm is analyzed. According to the distance between the prediction point of passenger and the real station and the weight of each prediction point, the accuracy of evaluation and prediction is evaluated and verified, which shows that the method is effective. According to the prediction model of arrival time, the actual bus arrival time is calculated. By comparing with the real value, the error is within a reasonable range. The database is established by using the calculation results of the average speed of the road section, and the road flow is classified and visualized in real time. The conclusion is in line with the actual state.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495;U491.17

【参考文献】