公交车到站时间预测模型与实证研究

发布时间：2018-06-25 08:06

本文选题：公交车到站时间 + 预测　；参考：《北京交通大学》2015年硕士论文

【摘要】：优先发展公共交通,是转变交通增长方式、提高资源利用效率、减少交通污染和缓解交通拥堵的有效方法,也是我国城市交通发展的主体思路。发展公共交通的关键在于为乘客提供良好的公共交通服务,公交车到站时间预测有利于出行者合理安排行程及公交车辆的实时调配,是提高公交服务水平的有效手段。论文实现了公交车辆运行的GPS位置数据与公交线路数据的匹配,通过插值得到公交车每秒的位置信息,且根据公交车到达站点的先后顺序,提取了前后车关系,并系统地分析了公交车的运行特性及公交车到站时间的影响因素。在此基础上,针对现有研究中很少明确地研究公交车在站点的停靠时间及公交车通过交叉口的时间,且较少考虑道路交通流状态对公交车运行时间的影响这两个问题,论文首先针对路段运行时间及车站停靠时间,提出了基于ε-支持向量机回归(ε-SVR)的无交叉口公交车到站时间预测模型,利用遗传算法优化模型参数及输入变量,并基于北京市三环300路内公交车的实例数据,分别按照工作日高峰(7：00～9：00/17：00～19：00)、平峰,以及周末3个时段,验证了模型的有效性。预测结果表明,通过引入路段检测器数据,能有效降低模型的预测误差。上述三个时段模型的平均相对预测误差最大分别能减少1.55%、1.50%和1.25%。论文进一步针对公交车通过交叉口的时间,提出了考虑交叉口的公交车到站时间预测模型,并基于VISSIM仿真,验证了模型的有效性。VISSIM仿真结果表明,对于路段的交通流处于畅通和缓行状态时,交叉口预测模型都能获得较高的预测精度,而对于拥挤状态,由于可能存在公交车在停车线前二次排队的现象,所以平均相对预测误差相对较大。
[Abstract]:Giving priority to the development of public transport is an effective way to change the way of traffic growth, improve the efficiency of resource utilization, reduce traffic pollution and alleviate traffic congestion. It is also the main idea of the development of urban traffic in China. The key to the development of public transportation is to provide good public service for passengers, and the prediction of bus arrival time is beneficial to travel. It is an effective way to improve the level of public transport service by arranging the itinerary and arranging the traffic in real time.
The paper realizes the matching between the GPS position data of the bus running and the bus line data, gets the position information of the bus per second through interpolation, and extracts the relationship between the front and back vehicles according to the order of the bus arrival site, and systematically analyzes the operating characteristics of the bus and the influencing factors of the bus arrival time.
On this basis, we seldom study the time of the stop time of the bus at the station and the time of the bus passing through the intersection in the existing research, and take less consideration to the two problems that the road traffic flow state affects the bus running time. First, the paper puts forward the epsilon support direction for the section running time and the station stop time. By using the genetic algorithm to optimize the model parameters and input variables of the bus arrival time without intersection (-SVR), and based on the example data of the three ring 300 road buses in Beijing, the validity of the model is verified according to the peak of working day (7:00 ~ 9:00/17:00 to 19:00), flat peak, and 3 periods of the weekend. The prediction results show that the prediction error of the model can be reduced effectively by introducing the link detector data. The maximum average relative prediction error of the three period models can be reduced by 1.55%, 1.50% and 1.25%. respectively.
The paper further aims at the time of the bus crossing through the intersection, and puts forward the bus arrival time prediction model considering the intersection. Based on the VISSIM simulation, the validity of the model is verified by.VISSIM simulation results. The results show that the intersection forecast model can obtain higher prediction precision for the traffic flow in the smooth and slow state of the section. For the crowded condition, the average relative prediction error is relatively large because there may be two queues in front of the parking line.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.17

【参考文献】