基于欠采样GPS数据的信号交叉口延误估计研究

发布时间：2018-01-13 04:28

本文关键词：基于欠采样GPS数据的信号交叉口延误估计研究　出处：《哈尔滨工业大学》2014年硕士论文　论文类型：学位论文

【摘要】：目前，，城市交通拥堵日益严重且多发生在信号交叉口，信号交叉口运行状态的良好与否直接关系到整个路网的运行效率与服务质量。信号控制交叉口服务质量与运行状态的评价主要取决于车辆通过信号交叉口时产生的延误。目前最为常用的延误计算方法仍然局限于传统的延误公式计算或者现场调查实测，不论何种方法都需要投入大量的人力和物力去对基础交通参数进行调查。随着GPS定位技术的应用，其为估算信号交叉口延误提供了一种新的高效方法，然而由于计算需要，现有方法需要1s采样间隔的GPS数据才能估计出信号交叉口延误，这样无疑增加了数据的采集、传输及存储成本，并且受采样频率限制，仍需自行采集。现如今我国绝大多数大中城市中，公交车、出租车均配备有GPS定位设备，由于数据只用于实现运营车辆的监控与调度，采样间隔多设定为30s，并且大量数据后期处于闲置状态。另外由于公交、出租车规模较大，几乎遍布整个路网，数据回传实时性强，因此有必要对此提出一种基于30s采样间隔，即欠采样GPS数据的信号交叉口延误估计方法。本文的重点是针对目前我国城市中现有的公交车、出租车等GPS数据欠采样特点，研究基于欠采样GPS数据估算信号交叉口延误的方法。首先对现有GPS数据计算信号交叉口控制延误进行了介绍，并对GPS数据的采样间隔进行分析，对GPS欠采样进行了界定，将采样间隔大于10s的GPS数据定义为欠采样GPS数据。同时对欠采样GPS数据反映出的车辆延误过程进行了分析，明确了利用欠采样GPS数据估算交叉口控制延误的关键在于对车辆延误过程进行轨迹还原，获得延误过程的关键时间节点，即初始减速时刻与加速截止时刻。本文采用对车辆在信号交叉口延误中加减速过程建模的方法还原车辆行驶轨迹，获取延误过程关键时间点。然后结合欠采样GPS数据特点，分有停车和无停车两种情况对车辆通过信号交叉口的加减速过程建立分段匀减速模型与分段匀加速模型，并通过对车辆在信号交叉口的加减速过程进行定性与定量的描述，验证了模型的可靠性。通过将带有约束条件的非线性非正定方程组问题转化为带有约束条件的非线性规划求最优解问题，对有停车过程的分段匀减速和匀加速模型给出了模型求解算法。对1s采样间隔的车辆通过信号交叉口的GPS数据以30s采样间隔进行重新采样，利用所建模型计算了车辆在信号交叉口的控制延误，计算精度大于85%，最后利用北京公交GPS数据进行实例分析，对闹市口东西直行方向的平均控制延误进行计算，并与Webster和HCM2010延误计算模型的计算结果进行对比分析，相对误差在20%以内，证明了本文研究方法的实用性。
[Abstract]:With the application of GPS positioning technology , it is necessary to input a large amount of manpower and material resources to survey the basic traffic parameters . With the application of GPS positioning technology , it provides a new efficient method for estimating the signalized intersection delay . With the application of GPS positioning technology , it is necessary to input the GPS data of 1s sampling interval to estimate the delay of signalized intersection , which undoubtedly increases the acquisition , transmission and storage cost of the data , and is limited by sampling frequency , and still needs to be collected automatically . In most Chinese cities , buses and taxis are equipped with GPS positioning equipment . Because the data is only used to realize the monitoring and dispatching of operating vehicles , the sampling interval is set to 30s , and a large amount of data is idle . In addition , it is necessary to propose a signal intersection delay estimation method based on 30s sampling interval , that is , under - sampled GPS data . This paper focuses on the method of estimating signalized intersection delay based on undersampled GPS data based on the characteristics of GPS data undersampling , such as buses , taxis and other GPS data currently in our country . Firstly , the delay of the control delay of GPS data is introduced and the sampling interval of GPS data is analyzed . The GPS data is defined as undersampled GPS data . At the same time , the delay process of vehicle caused by undersampling GPS data is analyzed . It is clear that the key point in estimating the delay of intersection control by undersampling GPS data is to get the key time node of the delay process , namely , the initial deceleration time and the accelerating cut - off time . In this paper , the vehicle running track is reduced by the method of modeling the delay of the signal intersection delay , and the key time point of the delay process is obtained . Then , by combining the characteristics of the undersampled GPS data , the vehicle passing through the deceleration process of the signal intersection is divided into a section uniform deceleration model and a segmented homogeneous acceleration model , and the reliability of the model is verified by the qualitative and quantitative description of the acceleration and deceleration process of the vehicle in the signal intersection . In this paper , the problem of nonlinear programming with constraint condition is transformed into nonlinear programming with constraint condition to solve the optimal solution problem , and the model solving algorithm is given to the section uniform deceleration and homogeneous acceleration model with the parking process . The GPS data of 1s sampling interval is re - sampled at 30 s sampling interval through GPS data at signalized intersection . The model is used to calculate the delay of the vehicle ' s control delay at signalized intersection , the calculation precision is more than 85 % , and finally , the average control delay in the direction of the downtown area is calculated and compared with the calculation result of Webster and HCM2010 delay calculation model . The relative error is within 20 % , which proves the practicability of the research method in this paper .

【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.23

【参考文献】