海量低频浮动车数据道路匹配及行程时间估算

发布时间：2018-04-20 15:38

本文选题：浮动车数据 + 海量　；参考：《武汉大学》2013年博士论文

【摘要】：道路交通信息的全面、准确、快速获取是城市交通管理、交通规划的基础,对缓解大城市交通拥堵,提供有效的大众出行指导具有重要的意义。浮动车是一种安装有全球定位设备并通过无线通讯系统将车辆状态和信息发送出的车辆,浮动车数据能及时准确的反映车辆所行驶道路的交通状况,是全面、快速获取道路交通信息的重要途径。本文以武汉市中上万辆出租车为载体的低频浮动车数据为研究对象,以准确、快速浮动车数据道路匹配和路段行程时间准确估算为目标,对大城市环境下海量低频浮动车数据的处理进行了讨论和研究,并通过真实、海量的浮动车数据和城市路网数据对研究的成果进行了验证。本文的主要的研究工作包括以下四个方面： 1、海量低频浮动车数据的分析与预处理。对武汉市一万二千多辆出租车获取的浮动车数据的格式、瞬时速度及航向、不同时段的数据量、采样时间间隔以及载客状态相关的数据和信息进行了分析。对于导航道路地图加密产生道路偏移的问题提出了一种道路地图栅格化的道路坐数据标系与浮动车数据坐标系的标定方法,提高了二种坐标系之间标定的精度,为海量浮动车数据的有效处理奠定了基础。 2、海量浮动车数据快速道路初匹配算法的研究海量浮动车数据道路匹配的算法效率是影响此类数据应用的重要因素。道路初匹配是根据匹配度计算将小于阈值的道路作为浮动车数据的候选匹配道路,由于一天的浮动车数据就有约一万四千个,并且道路的数量有二万六千多条,算法的效率对浮动车处理的时间影响极大。本文首先对基于地图格网化的浮动车数据道路匹配算法进行了分析和讨论,给出了一种面向计算效率的地图格网划分最佳参数；还提出了一种基于道路地图栅格化的海量浮动车数据地图初匹配的算法,使得海量浮动车数据完成道路初匹配计算的时间更短。 3、基于序列低频浮动车数据路径计算的研究首先研究了基于序列浮动车数据路径计算中一次路径计算选取浮动车数据点数量的问题,指出了在城市复杂路网条件下路径计算点只有达到或超过3个才可能保障路径重建正确,同时路径计算点越多,重建结果可靠性越大,但计算的复杂度也越大。然后针对相关的浮动车数据质量不稳定,载客状态发生变化等情况进行了研究,并对算法进行了优化改进,提高了计算结果的准确性。还研究了路径算法中路径搜索区域道路端点的甄选方法,的提高路径计算效率的方法,减少了路径计算的时间。 4、基于低频浮动车数据路段行程时间估算的研究路段行程时间是道路交通中的一项最重要的参数。在分析已有利用低频浮动车数据进行路段行程时间计算算法的基础上提出了一种基于道路交叉口下游路段浮动车数据的路段行程时间估算方法,在分析低频浮动车数据在道路交叉口附近路段上的位置、速度信息,结合车辆在道路交叉口附近路段行驶的特点,给出了一个车辆通过道路交叉口时刻的计算模型。基于这个模型,在获得道路交叉口下游路段上低频浮动车数据的位置和速度信息后,能较准确地计算出浮动车通过路口的时刻,通过对路段上下游二个路口浮动车通过时刻的计算,从而能较准确地估算出路段的通行时间。
[Abstract]:The comprehensive, accurate and rapid access to road traffic information is the basis of urban traffic management and traffic planning. It is of great significance to alleviate traffic congestion in large cities and provide effective guidance for mass travel. The floating car is a vehicle installed with global positioning equipment and sent through the wireless communication system to send the vehicle status and information. The vehicle data can reflect the traffic condition of the road in time and accurately. It is an important way to obtain the road traffic information in an all-round way. This paper takes the data of the low frequency floating car as the research object, which is the carrier of the Wuhan city. The target is accurate, the data Road matching of the fast floating car and the accurate estimate of the road travel time are the target. The data processing of massive low frequency floating car in the large city environment is discussed and studied, and the results of the research are verified through the real, massive floating car data and urban road network data. The main research work of this paper includes the following four aspects:
1, analysis and preprocessing of the data of the mass low frequency floating car. The data and information of the floating car data obtained by more than 12000 taxis in Wuhan are analyzed, such as the data of the instantaneous speed and course, the amount of data in different periods, the interval of sampling time, and the data and information related to the state of the passengers. The problem puts forward a method to calibrate the road map grid and the data coordinate system of the floating car, which improves the accuracy of the calibration between the two coordinate systems, and lays the foundation for the effective processing of the data of the mass floating car.
2, the research on the algorithm for the initial matching of the mass floating vehicle data fast path matching algorithm is an important factor affecting the data application of the floating car. The initial road matching is the candidate matching path of the floating car data based on the matching degree calculation, which is less than the threshold, because the floating car data of one day is about one. There are all four thousand, and there are more than 26000 roads. The efficiency of the algorithm has a great influence on the time of the floating car. Firstly, the paper analyzes and discusses the data road matching algorithm based on the map grid, and gives the best parameter of the geo grid, which is oriented to the calculation efficiency. The initial matching algorithm of massive floating car data rasterization on road map rasterization makes the data of mass floating vehicle data shorter than that of road initial matching.
3, based on the study of the data path calculation of the sequence low frequency floating car, the problem of selecting the number of data points of the floating car based on the single path calculation in the data path calculation of the serial floating car is first studied. It is pointed out that only the path calculation point is only up to 3 or more in the complex road network conditions. The more calculation points, the greater the reliability of the reconstruction results, the greater the complexity of the calculation, and then research on the unstable data quality of the related floating car, the change of the passenger state and so on, and optimize the algorithm to improve the accuracy of the calculation results. The selection method and the method of improving the calculation efficiency of the path reduce the computation time of the path.
4, based on the estimation of the travel time of the data section of the low frequency floating vehicle, the travel time of the section is one of the most important parameters in the road traffic. On the basis of the analysis of the calculation algorithm of the link travel time of the low frequency floating car, a link trip based on the floating vehicle data of the lower section of the road intersection is proposed. In the analysis of the location of the low frequency floating car data near the road intersection, the velocity information and the characteristics of the vehicle running at the road intersection, the calculation model of a vehicle passing through the road intersection is given. Based on this model, a low frequency floating car on the lower section of the road intersection is obtained. After the location and speed information of the data, it can accurately calculate the time of the floating train passing through the intersection. Through the calculation of the passing time of the floating car at the two intersection of the upper and lower reaches of the section, the time of the section can be accurately estimated.

【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：U495;P208

【参考文献】