基于路口相关性的变结构式组合交通流量预测的研究

发布时间：2018-03-24 19:24

本文选题：智能交通　切入点：交通流量　出处：《合肥工业大学》2017年硕士论文

【摘要】：随着智能交通系统的快速发展,智能交通的实时、准确、高效的控制俨然已成为智能交通系统中的重要部分。交通流量预测作为智能交通系统的基础,可对交通诱导,道路拥堵预警以及最优路径的选择提供有效的参考数据。预测的精确性直接影响交通管理和控制的成效。同时,由于道路拥堵现象的存在,选择出合适的驾驶路径,节约驾驶时间也显得格外重要。目前,汽车导航越来越普及,但并不能为人类智能的选择出耗时最短的路径,因此,提供最优行驶路径对交通诱导意义重大。针对提高短时交通流量预测模型的精度以及准确性,本文首先对采集的原始交通流量进行预处理,完成原始交通流量数据归一化、简约、修正和补全后,在小波分析的基础上,把交通流量序列信号分解为低频序列信号和高频序列信号;接着依据卡尔曼滤波模型在处理平稳数据上的优势对原始交通流量的低频序列信号进行预测和RBF神经网络的高动态非线性映射性对高频序列信号进行预测,建立了一种变结构式组合交通流量预测模型,得出预测结果,并对预测误差进行分析;最后在预测结果以及车辆导航系统数据处理算法分析的基础上,运用Vissim交通控制软件赋值在模拟路网,采用Floyd算法选择出基于时间权重的路径,结合变结构式组合交通流量预测模型的预测算法,利用VS(Microsoft Visual Studio)实现软件功能的开发。
[Abstract]:With the rapid development of intelligent transportation system, real-time, accurate and efficient control of intelligent transportation has become an important part of intelligent transportation system. As the basis of intelligent transportation system, traffic flow forecasting can guide traffic. The prediction accuracy directly affects the effectiveness of traffic management and control. At the same time, due to the existence of traffic congestion, the appropriate driving path is chosen. It is especially important to save driving time. At present, car navigation is becoming more and more popular, but it can not choose the shortest path for human intelligence, so, In order to improve the accuracy and accuracy of the short-term traffic flow prediction model, this paper first preprocesses the collected original traffic flow and completes the normalization of the original traffic flow data. On the basis of wavelet analysis, the traffic flow sequence signal is decomposed into low frequency sequence signal and high frequency sequence signal. Then according to the advantage of Kalman filter model in processing stationary data, the low frequency sequence signal of original traffic flow is predicted and the high dynamic nonlinear mapping of RBF neural network is used to predict the high frequency sequence signal. A variable structure integrated traffic flow forecasting model is established, the prediction results are obtained, and the prediction error is analyzed. Finally, based on the analysis of the prediction results and the data processing algorithm of the vehicle navigation system, The Vissim traffic control software assignment is used to simulate the road network, the Floyd algorithm is used to select the path based on time weight, the variable structure combined traffic flow forecasting algorithm is combined, and the software function is developed by VS(Microsoft Visual Studio.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.14

【参考文献】