城市快速路路段行程时间估计与预测方法研究

发布时间：2019-06-10 06:05

【摘要】：路段行程时间是描述道路交通状态的重要参数,它能够较好地评价道路的通畅程度,能够反映道路的运输效率,在交通规划、交通管理与交通控制中起着重要作用,在当前智能交通系统的研究和开发应用中也占据着重要地位。针对城市快速路的路段行程时间估计问题,考虑到微波检测器技术成熟、数据易获取以及低成本的特点,本文提出了一种基于微波检测数据的行程—时间域法进行路段行程时间估计。该方法首先假定微波检测器实时检测的速度即为路段单元在不同时间单元的空间平均车速,然后构建车辆出行的行程—时间域,最后通过模拟虚拟车辆穿越行程—时间域的过程获得车辆在该路段上的行程时间。该方法以北京市二环快速路上的微波检测数据为基础进行实例验证,结果表明,相比于传统静态行程时间估计方法,该方法显著提高了行程时间估计精度。不仅获得当前时刻的路段行程时间非常重要,预测未来时刻的路段行程时间也十分重要。本文以提高路段行程时间预测精度为目的,构建了基于小波神经网络的路段行程时间预测模型。然后以行程—时间域法估计得到的北京市二环快速路路段行程时间为实验数据,根据不同参数选择、不同样本数据建立多个预测实例对该模型进行检验,并与BP神经网络模型的预测误差进行比较。结果分析表明,所建立的小波神经网络模型能够更好地描述输入输出的映射规律。最后,将各个预测实例的结果进行对比,结合以往的路段行程时间预测研究,进一步分析了误差产生的原因以及本文所构建的模型取得较高预测精度的原因。本文所构建的路段行程时间预测模型及对模型进行的相关讨论,对于交通参数预测领域的研究具有一定的创新意义和借鉴价值。
[Abstract]:The travel time of road section is an important parameter to describe the state of road traffic. It can better evaluate the unobstructed degree of the road, can reflect the transportation efficiency of the road, and plays an important role in traffic planning, traffic management and traffic control. It also occupies an important position in the research, development and application of intelligent transportation system. In view of the problem of road travel time estimation of urban expressway, considering the mature technology of microwave detector, easy to obtain data and low cost, In this paper, a travel-time domain method based on microwave detection data is proposed to estimate the travel time of road sections. The method first assumes that the speed detected by the microwave detector in real time is the spatial average speed of the section unit in different time units, and then constructs the travel-time domain of the vehicle. Finally, the travel time of the virtual vehicle on the road section is obtained by simulating the process of crossing the travel-time domain of the virtual vehicle. The method is verified by an example based on the microwave detection data on the second Ring Road Expressway in Beijing. The results show that compared with the traditional static travel time estimation method, this method significantly improves the accuracy of travel time estimation. It is very important not only to obtain the travel time of the current time, but also to predict the travel time of the road section in the future. In order to improve the accuracy of road travel time prediction, a wavelet neural network based travel time prediction model is constructed in this paper. Then, taking the travel time estimated by the travel-time domain method as the experimental data, several prediction examples are established to test the model according to the selection of different parameters and different sample data. The prediction error is compared with that of BP neural network model. The results show that the wavelet neural network model can better describe the mapping law of input and output. Finally, the results of each prediction example are compared, and combined with the previous research on road travel time prediction, the causes of errors and the reasons for the higher prediction accuracy of the model constructed in this paper are further analyzed. The road travel time prediction model constructed in this paper and the related discussion of the model have certain innovative significance and reference value for the research in the field of traffic parameter prediction.
【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】：U491.14

【参考文献】