自由式路网交通拥堵预报方法研究
发布时间：2018-10-12 12:22
【摘要】：随着经济的不断发展和日益加剧的城市化进程,汽车保有量逐年增加,现代城市交通面临越来越多的问题,如交通拥堵、交通污染和交通事故等。这些问题对人们的出行造成许多影响,降低人们出行效率,使城市道路面临越来越大的交通压力。自由式路网城市由于地形限制,交通现象尤为复杂,城市道路交通拥堵严重,拥堵疏散较为困难,为缓解自由式路网交通拥堵,对其交通流参数进行短时预测,并根据预测结果对交通拥堵状况进行判定识别,发出实时的交通拥堵预报信息非常重要。本论文以自由式路网交通流为研究对象,介绍交通流基本特征参数,分析交通流参数的采集技术及其适用范围,研究交通信息与处理方法。研究自由式路网交通拥堵的交通流特性和时空分布特征,对自由式路网交通拥堵原因进行深入分析。对自由式路网来说,交通流参数的相关度和城市道路断面之间的相对位置有关,距离较远,相关性比较低,同样距离近的道路断面由于某些道路交通控制措施(如单向交通或交叉口禁左等)相关性不一定强。据此,本文提出基于空间相关性的自由式路网交通参数短时预测模型。通过分析和比较单一预测方法的优缺点和适用范围,确定采用基于多维新息算法的ARIMA预测方法和RBF神经网络预测方法对自由式路网交通参数进行短时预测,并将这两种预测方法根据最优权分配进行组合,利用组合模型对交通参数进行短时预测,并通过算例验证本论文研究的组合模型对交通参数短时预测效果。根据路网交通流分布的时间和空间特性,本文将自由式路网交通拥堵状况分为四个等级,即畅通、较拥堵、拥堵、堵塞。分别从交叉口、路段两方面选定评价指标,形成指标体系。采用基于层次熵定权的模糊综合评判模型对预测得到的交通流参数进行评价,识别交通拥堵状况,进而发出交通拥堵实时预报信息。本文提出的交通拥堵预报方法可以及时预报交通拥堵的发生,对交通参与者交通出行路线的选择和交通管理者对交通拥堵疏散决策均有一定的参考价值。
[Abstract]:With the continuous development of economy and the increasing process of urbanization, the number of cars is increasing year by year. Modern urban traffic is facing more and more problems, such as traffic congestion, traffic pollution and traffic accidents. These problems have a lot of impact on people's travel, reduce people's travel efficiency, and make urban roads face more and more traffic pressure. Because of the terrain restriction, the traffic phenomenon is especially complex in the free-type road network city. The traffic congestion is serious and the evacuation is difficult. In order to alleviate the traffic congestion of the free road network, the traffic flow parameters of the freeway road network are forecasted in a short time. According to the prediction results, it is very important to identify the traffic congestion and send out real-time traffic congestion forecast information. In this paper, the basic characteristic parameters of traffic flow are introduced, the collection technology of traffic flow parameters and its applicable scope are analyzed, and the traffic information and processing methods are studied. The characteristics of traffic flow and space-time distribution of free road network traffic congestion are studied, and the causes of free road network traffic congestion are analyzed. For free road networks, the correlation of traffic flow parameters is related to the relative position between urban road sections, the distance is relatively long, the correlation is relatively low. Because of some road traffic control measures (such as one-way traffic or intersection forbidden left), the correlation of the same close road section is not necessarily strong. On the basis of this, this paper presents a short-time prediction model of traffic parameters of freestyle road network based on spatial correlation. By analyzing and comparing the advantages and disadvantages of the single forecasting method and its application scope, the ARIMA forecasting method based on the multidimensional innovation algorithm and the RBF neural network forecasting method are adopted to predict the traffic parameters of the freestyle road network in a short time. The two forecasting methods are combined according to the optimal weight allocation, and the combined model is used to predict the traffic parameters in a short time, and an example is given to verify the effect of the combined model on the short-term prediction of traffic parameters. According to the time and space characteristics of road network traffic flow distribution, this paper divides the free road network traffic congestion into four grades, that is, unblocked, more congested, and blocked. The evaluation index is selected from two aspects of intersection and section, and the index system is formed. The fuzzy comprehensive evaluation model based on hierarchical entropy weight is used to evaluate the traffic flow parameters, identify the traffic congestion, and send out the real-time traffic congestion forecast information. The traffic congestion forecasting method proposed in this paper can forecast the occurrence of traffic congestion in time. It has certain reference value for traffic participants' route choice and traffic manager's decision of traffic congestion evacuation.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491
本文编号：2266096
[Abstract]:With the continuous development of economy and the increasing process of urbanization, the number of cars is increasing year by year. Modern urban traffic is facing more and more problems, such as traffic congestion, traffic pollution and traffic accidents. These problems have a lot of impact on people's travel, reduce people's travel efficiency, and make urban roads face more and more traffic pressure. Because of the terrain restriction, the traffic phenomenon is especially complex in the free-type road network city. The traffic congestion is serious and the evacuation is difficult. In order to alleviate the traffic congestion of the free road network, the traffic flow parameters of the freeway road network are forecasted in a short time. According to the prediction results, it is very important to identify the traffic congestion and send out real-time traffic congestion forecast information. In this paper, the basic characteristic parameters of traffic flow are introduced, the collection technology of traffic flow parameters and its applicable scope are analyzed, and the traffic information and processing methods are studied. The characteristics of traffic flow and space-time distribution of free road network traffic congestion are studied, and the causes of free road network traffic congestion are analyzed. For free road networks, the correlation of traffic flow parameters is related to the relative position between urban road sections, the distance is relatively long, the correlation is relatively low. Because of some road traffic control measures (such as one-way traffic or intersection forbidden left), the correlation of the same close road section is not necessarily strong. On the basis of this, this paper presents a short-time prediction model of traffic parameters of freestyle road network based on spatial correlation. By analyzing and comparing the advantages and disadvantages of the single forecasting method and its application scope, the ARIMA forecasting method based on the multidimensional innovation algorithm and the RBF neural network forecasting method are adopted to predict the traffic parameters of the freestyle road network in a short time. The two forecasting methods are combined according to the optimal weight allocation, and the combined model is used to predict the traffic parameters in a short time, and an example is given to verify the effect of the combined model on the short-term prediction of traffic parameters. According to the time and space characteristics of road network traffic flow distribution, this paper divides the free road network traffic congestion into four grades, that is, unblocked, more congested, and blocked. The evaluation index is selected from two aspects of intersection and section, and the index system is formed. The fuzzy comprehensive evaluation model based on hierarchical entropy weight is used to evaluate the traffic flow parameters, identify the traffic congestion, and send out the real-time traffic congestion forecast information. The traffic congestion forecasting method proposed in this paper can forecast the occurrence of traffic congestion in time. It has certain reference value for traffic participants' route choice and traffic manager's decision of traffic congestion evacuation.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491
【参考文献】
相关期刊论文 前4条
1 黄大荣;宋军;汪达成;曹建秋;李伟;;基于ARMA和小波变换的交通流预测模型研究[J];计算机工程与应用;2006年36期
2 温惠英;周玮明;;基于S型函数预处理的RBF神经网络交通流量预测[J];交通信息与安全;2009年04期
3 张晓利;陆化普;;非参数回归方法在短时交通流预测中的应用[J];清华大学学报(自然科学版)网络.预览;2009年09期
4 朱中,杨兆升;实时交通流量人工神经网络预测模型[J];中国公路学报;1998年04期
相关博士学位论文 前1条
1 王宏涛;径向基函数和神经网络技术在逆向工程中的应用研究[D];南京航空航天大学;2006年
相关硕士学位论文 前1条
1 杨俊;基于核主成分分析和径向基神经网络的文本分类研究[D];中国科学技术大学;2009年
，本文编号：2266096
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2266096.html
