城市交通流短时预测模型研究

发布时间：2018-08-03 18:29

【摘要】：随着经济水平的提高,家庭负担的减少,私家车的购买量日益激增,交通拥堵是困扰着中国乃至当今国际的一大交通难题,如何缓解交通的压力成为我国亟待解决的问题。获得实时而又准确的交通流量是进行交通诱导和控制的基础,更是解决各种交通问题的关键所在。首先本文在对短时交通流预测的国内外研究现状进行分析的基础上,对城市交通流的特性进行了分析,总结了现有的预测方法,提出了基于相空间重构的卡尔曼滤波交通流预测仿真模型。为得到隐藏在短时交通流一维时间序列中的特性,对一维的时间序列进行重构,利用C?C算法来确定空间重构的延迟时间和嵌入维数,利用相空间重构得到的相点来描述由状态向量构成相点的状态空间,再结合卡尔曼滤波理论对实测数据进行下一时刻的预测并校正相点的未来的发展规律,在这两种理论的基础上建立短时交通流的预测模型,最后根据某路段的实际交通情况进行仿真验证。其次,对支持向量机SVM理论进行细致的研究分析,并针对本文的预测对象确定核函数的种类,针对方法的不足在数据训练之前引入小波去噪理论,根据几种小波的特点,选用加入折中因子的阈值去噪法对数据进行去噪,同时为了提高预测的准确性通过蚁群优化的算法对模型的参数进行优化,构建了基于参数优化的SVM的预测模型,并结合实际的交通流量进行仿真分析,验证了基于该算法的可用性和实用性。最后为了对比分析,分别将构建的相空间重构的卡尔曼滤波交通流预测仿真模型和基于参数优化的SVM短时交通流预测模型进行实验仿真,并利用文中提出的指标信息对两组数据的预测结果进行综合对比,仿真结果表明基于智能算法的参数优化的SVM模型更有效的提高了交通流的预测精度,证明这种智能组合算法能取得更好的预测效果。
[Abstract]:With the improvement of economic level, the reduction of family burden and the increasing purchase of private cars, traffic congestion is a major transportation problem that puzzles China and even the international community. How to relieve the pressure of traffic becomes an urgent problem to be solved in our country. Obtaining real-time and accurate traffic flow is the basis of traffic guidance and control, and the key to solve various traffic problems. Firstly, based on the analysis of the current situation of short-term traffic flow prediction at home and abroad, the characteristics of urban traffic flow are analyzed, and the existing forecasting methods are summarized. A simulation model of Kalman filter traffic flow prediction based on phase space reconstruction is proposed. In order to obtain the characteristics hidden in the one-dimensional time series of short-time traffic flow, the one-dimensional time series is reconstructed, and the delay time and embedding dimension of the spatial reconstruction are determined by using CnC algorithm. The phase points obtained by phase space reconstruction are used to describe the state space which is composed of state vectors, and then the prediction of the next moment of the measured data and the correction of the future development law of the phase points are carried out based on the Kalman filter theory. On the basis of these two theories, the short-term traffic flow prediction model is established, and the simulation is carried out according to the actual traffic situation of a certain section of the road. Secondly, the support vector machine (SVM) SVM theory is studied and analyzed in detail, and the kernel function is determined according to the prediction object in this paper. The wavelet denoising theory is introduced before the data training to overcome the shortcomings of the method, according to the characteristics of several kinds of wavelets. In order to improve the accuracy of prediction, the SVM prediction model based on parameter optimization is constructed in order to improve the accuracy of prediction and optimize the parameters of the model by means of ant colony optimization algorithm. The availability and practicability of the algorithm are verified by simulation and analysis of actual traffic flow. Finally, in order to compare and analyze, the constructed Kalman filter traffic flow prediction simulation model based on phase space reconstruction and the SVM short-time traffic flow prediction model based on parameter optimization are simulated. The simulation results show that the parameter optimization SVM model based on intelligent algorithm can improve the prediction accuracy of traffic flow more effectively. It is proved that this intelligent combination algorithm can achieve better prediction results.
【学位授予单位】：河南理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.14;TP18

【相似文献】