【摘要】：隨著經(jīng)濟的不斷發(fā)展和日益加劇的城市化進程,汽車保有量逐年增加,現(xiàn)代城市交通面臨越來越多的問題,如交通擁堵、交通污染和交通事故等。這些問題對人們的出行造成許多影響,降低人們出行效率,使城市道路面臨越來越大的交通壓力。自由式路網(wǎng)城市由于地形限制,交通現(xiàn)象尤為復(fù)雜,城市道路交通擁堵嚴(yán)重,擁堵疏散較為困難,為緩解自由式路網(wǎng)交通擁堵,對其交通流參數(shù)進行短時預(yù)測,并根據(jù)預(yù)測結(jié)果對交通擁堵狀況進行判定識別,發(fā)出實時的交通擁堵預(yù)報信息非常重要。本論文以自由式路網(wǎng)交通流為研究對象,介紹交通流基本特征參數(shù),分析交通流參數(shù)的采集技術(shù)及其適用范圍,研究交通信息與處理方法。研究自由式路網(wǎng)交通擁堵的交通流特性和時空分布特征,對自由式路網(wǎng)交通擁堵原因進行深入分析。對自由式路網(wǎng)來說,交通流參數(shù)的相關(guān)度和城市道路斷面之間的相對位置有關(guān),距離較遠,相關(guān)性比較低,同樣距離近的道路斷面由于某些道路交通控制措施(如單向交通或交叉口禁左等)相關(guān)性不一定強。據(jù)此,本文提出基于空間相關(guān)性的自由式路網(wǎng)交通參數(shù)短時預(yù)測模型。通過分析和比較單一預(yù)測方法的優(yōu)缺點和適用范圍,確定采用基于多維新息算法的ARIMA預(yù)測方法和RBF神經(jīng)網(wǎng)絡(luò)預(yù)測方法對自由式路網(wǎng)交通參數(shù)進行短時預(yù)測,并將這兩種預(yù)測方法根據(jù)最優(yōu)權(quán)分配進行組合,利用組合模型對交通參數(shù)進行短時預(yù)測,并通過算例驗證本論文研究的組合模型對交通參數(shù)短時預(yù)測效果。根據(jù)路網(wǎng)交通流分布的時間和空間特性,本文將自由式路網(wǎng)交通擁堵狀況分為四個等級,即暢通、較擁堵、擁堵、堵塞。分別從交叉口、路段兩方面選定評價指標(biāo),形成指標(biāo)體系。采用基于層次熵定權(quán)的模糊綜合評判模型對預(yù)測得到的交通流參數(shù)進行評價,識別交通擁堵狀況,進而發(fā)出交通擁堵實時預(yù)報信息。本文提出的交通擁堵預(yù)報方法可以及時預(yù)報交通擁堵的發(fā)生,對交通參與者交通出行路線的選擇和交通管理者對交通擁堵疏散決策均有一定的參考價值。
[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.
【學(xué)位授予單位】：重慶交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：U491

【參考文獻】

相關(guān)期刊論文 前4條

1 黃大榮;宋軍;汪達成;曹建秋;李偉;;基于ARMA和小波變換的交通流預(yù)測模型研究[J];計算機工程與應(yīng)用;2006年36期

2 溫惠英;周瑋明;;基于S型函數(shù)預(yù)處理的RBF神經(jīng)網(wǎng)絡(luò)交通流量預(yù)測[J];交通信息與安全;2009年04期

3 張曉利;陸化普;;非參數(shù)回歸方法在短時交通流預(yù)測中的應(yīng)用[J];清華大學(xué)學(xué)報(自然科學(xué)版)網(wǎng)絡(luò).預(yù)覽;2009年09期

4 朱中,楊兆升;實時交通流量人工神經(jīng)網(wǎng)絡(luò)預(yù)測模型[J];中國公路學(xué)報;1998年04期

相關(guān)博士學(xué)位論文 前1條

1 王宏濤;徑向基函數(shù)和神經(jīng)網(wǎng)絡(luò)技術(shù)在逆向工程中的應(yīng)用研究[D];南京航空航天大學(xué);2006年

相關(guān)碩士學(xué)位論文 前1條

1 楊俊;基于核主成分分析和徑向基神經(jīng)網(wǎng)絡(luò)的文本分類研究[D];中國科學(xué)技術(shù)大學(xué);2009年

，

本文編號：2266096