基于时空轨迹数据的交通状态分析及预测

发布时间：2019-01-24 08:41

【摘要】：交通压力已经成为每个城市发展到一定程度必然面临的难题,缓解交通压力需要科学有效的交通管理措施。采取何种方式获取交通路网中的信息和数据,实现对城市路网的交通状态的实时监测,并将路网中海量的实时交通信息与庞大的历史数据库相结合,选取合适的交通参数,建立有效的数学模型,及时准确地分析出道路交通状态的变化,进而预测未来时段的交通情况,为交通信息服务、交通控制与诱导提供帮助,是智能交通系统的关键内容。本文选用出行率高、路网覆盖率高的出租车GPS数据作为优质的浮动车时空轨迹数据实时反映城市路网的交通状况,利用基于Hadoop平台的大数据技术管理和处理庞大的交通状态信息,解决了传统方式在海量数据处理方面的难题,具有高效率、高准确性和高时效性的优点,这对及时分析交通状态并作出预测十分重要。同时,由于不同道路之间的差异所导致的复杂情况,仅以交通流量或行程时间等参数来描述道路的交通状况显得并不合适。为了更有效、准确、及时地分析和预测道路交通状态,本文选用路段速度作为更加合理的交通状态参数。此外,本文分析了路网划分后路段速度的时间序列,利用四分位数的特性优化算法,提高了路段速度模型的合理性和准确性,并通过真实历史数据验证四分位法的有效性。结果表明该方法既体现出了路段速度的变化趋势,同时削弱了极端值和异常值的影响,能够展现出合理的交通状态变化过程,并且其计算简便,为大规模数据处理有效地节省了计算资源。对计算结果的曲线拟合也证明了四分位法处理路段速度的可靠性。然后利用基于同比历史数据加权均值和修正值的数学模型,在划分路段上对未来的路段交通速度进行预测。结果表明该预测模型可以有效地预测出交通状态的变化趋势,路段速度的预测结果十分接近真实值。通过预测带来的结果,能够有效地帮助交通诱导服务和交通管理决策。
[Abstract]:Traffic pressure has become a difficult problem for every city to develop to a certain degree. To alleviate traffic pressure, scientific and effective traffic management measures are needed. How to obtain the information and data in the traffic network, realize the real-time monitoring of the traffic state of the urban road network, and combine the massive real-time traffic information in the road network with the huge historical database, and select the appropriate traffic parameters. It is the key content of Intelligent Transportation system to establish an effective mathematical model to analyze the change of road traffic state in time and accurately, and then to predict the traffic situation in the future, and to provide help for traffic information service, traffic control and guidance. In this paper, taxi GPS data with high trip rate and high network coverage are selected as the high quality floating vehicle space-time track data to reflect the traffic situation of the urban road network in real time. Using big data technology based on Hadoop platform to manage and deal with huge traffic state information, it solves the problem of mass data processing in traditional way, and has the advantages of high efficiency, high accuracy and high timeliness. This is very important for analyzing traffic conditions and making predictions in time. At the same time, because of the complexity caused by the difference between different roads, it is not appropriate to describe the road traffic condition only by the parameters such as traffic flow or travel time. In order to analyze and predict the state of road traffic more effectively, accurately and in time, this paper chooses the speed of road section as a more reasonable parameter of traffic state. In addition, this paper analyzes the time series of road speed after road network division, using the quartile characteristic optimization algorithm to improve the rationality and accuracy of the section speed model, and verifies the validity of the quartile method through the real historical data. The results show that the method not only reflects the changing trend of road speed, but also weakens the influence of extreme value and abnormal value, and it can show a reasonable process of traffic state change, and its calculation is simple and convenient. Computing resources are saved effectively for large-scale data processing. The curve fitting of the calculated results also proves the reliability of the quartile method. Then using the mathematical model based on the weighted mean value and the revised value of the historical data of the same year, the traffic speed of the future road section is forecasted in the division section. The results show that the prediction model can effectively predict the changing trend of traffic state, and the prediction result of road speed is very close to the real value. The prediction results can effectively help traffic guidance services and traffic management decisions.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491;TP311.13

【参考文献】