基于GPS数据的公交车停靠站时间预测研究

发布时间：2018-09-10 16:11

【摘要】：公交车辆的停靠站时间是影响公交运行效率的重要因素,对公交停靠站时间规律的了解,能够使公交系统管理者更准确地掌握公交车辆的运行规律,为公交站点的优化设置和公交动态调度提供理论支持。与此同时,由于公交停靠站时间是公交运行时间的重要组成部分,对公交停靠站时间的预测能够提高公交到站时间预测的精度,并且为公交优先信号控制等提供技术支持,从而提高公交系统的运营效率和服务水平。因此,对公交停靠站时间的研究具有重要的理论意义和现实需求。国内外对停靠站时间的研究主要以减少停靠站时间延误或优化公交站点设置为主,有关公交车停靠站时间的规律分析及预测方法的研究较少。本文在研究了公交车辆的停靠站时间的采集方法的基础上,分析公交车停靠站时间的时间相关性、空间相关性以及线路特性等规律,进而运用指数平滑等思想提出了基于时间序列的公交停靠站时间预测方法。本文的研究工作主要包括以下三方面内容:①分析了公交车停靠站时间的影响因素。公交站点作为城市公交线路的重要节点,公交车辆在其停靠所花费的时间受很多因素影响。本文根据所获取停靠站数据的特点,将公交停靠站时间分为减速进站时间、停车服务时间和加速出站时间三部分,并分别从进出站时间和停车服务时间两个角度分析其影响因素。②研究了公交车停靠站时间的相关规律。首先分析了停靠站时间的分布特征函数;其次,从时间和空间两方面研究了公交车辆停靠站时间的差异性和相关性,为后文中解决停靠站时间预测的关键问题提供相应支持;最后,本文分析了不同线路的公交车辆的停靠站时间特性,为建立停靠站时间预测模型奠定基础。③建立了基于时间序列的公交停靠站时间预测模型。针对GPS数据的特点及公交车停靠站时间的相关规律,按照线路、站点、时段及工作日四个维度建立公交车停靠站时间历史统计值,利用指数平滑思想,采用时间序列方法建立公交停靠站时间预测模型。之后利用实际数据对模型的预测精度进行检验,试验结果表明该模型具有较好的预测效果,能够比较准确地预测公交车辆的停靠站时间。本文基于大量数据全面分析了公交停靠站时间的时空规律和线路规律的分析,建立的基于时间序列的公交停靠站时间预测模型适用于大部分公交线路,经过试验验证预测精度较高,能够对公交停靠站时间的进一步研究起到积极的推动作用。
[Abstract]:The bus stop time is an important factor affecting the bus operation efficiency. Understanding the bus stop time regularity can help the bus system administrators to grasp the bus operation regularity more accurately and provide theoretical support for the optimal setting of bus stops and the dynamic dispatching of buses. As an important part of transit running time, the prediction of bus stop time can improve the accuracy of bus stop time prediction, and provide technical support for bus priority signal control, so as to improve the operational efficiency and service level of the bus system. Domestic and foreign researches on the bus stop time mainly focus on reducing the delay of bus stop time or optimizing the bus stop settings. There are few researches on the regularity analysis and prediction methods of bus stop time. Based on the time series theory, this paper proposes a time series prediction method for bus stops. The main research work of this paper includes the following three aspects: 1. The influencing factors of bus stops time are analyzed. According to the characteristics of the bus stop data obtained, the bus stop time is divided into three parts: deceleration arrival time, parking service time and acceleration departure time, and is analyzed from the angle of arrival and departure time and parking service time respectively. Secondly, the difference and correlation of bus stopping time are studied from both time and space, which provides the corresponding support for solving the key problems of bus stopping time prediction in the future. This paper analyzes the time characteristics of bus stops on different routes and lays a foundation for establishing the time prediction model of bus stops. Thirdly, the time prediction model of bus stops based on time series is established. This paper establishes the historical statistic value of bus stop time, uses the exponential smoothing thought, and uses the time series method to establish the prediction model of bus stop time. Then the prediction accuracy of the model is checked by the actual data. The experimental results show that the model has good prediction effect and can predict the bus stop more accurately. Based on a large number of data, this paper comprehensively analyzes the time-space rule and the line rule of bus stopping time. The time-series-based prediction model of bus stopping time is suitable for most bus routes. It has been verified by experiments that the prediction accuracy is high and can be used for further study of bus stopping time. Great impetus.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.17

【相似文献】