【摘要】：提高高速公路事故管理水平對(duì)于保障國(guó)民經(jīng)濟(jì)健康發(fā)展、人民出行安全舒適、社會(huì)生活平安和諧具有重要意義。本文首先從“危機(jī)管理”的角度對(duì)交通事故管理所涉及的各個(gè)階段進(jìn)行了分析，提出了高速公路交通事故管理系統(tǒng)的框架結(jié)構(gòu)并對(duì)框架各部分功能進(jìn)行了概述，總結(jié)了與各個(gè)基本功能部分有關(guān)的研究現(xiàn)狀，并選擇當(dāng)前研究比較薄弱的“事故事前處理”部分作為本文研究的重點(diǎn)，按交通流預(yù)測(cè)—交通事故預(yù)警—事故提前應(yīng)對(duì)的邏輯層次展開(kāi)研究，主要研究?jī)?nèi)容和結(jié)果包括： 一：交通流短時(shí)預(yù)測(cè)方法和預(yù)測(cè)系統(tǒng)研究，首先提出了一種基于內(nèi)聚度指標(biāo)和遺傳算法的狀態(tài)向量選擇算法，在四種不同的情景下對(duì)算法的效果進(jìn)行了驗(yàn)證，證明本文提出的算法能夠明顯提高非參數(shù)回歸預(yù)測(cè)和神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)的精度。隨后提出了一種數(shù)據(jù)驅(qū)動(dòng)的交通流短時(shí)預(yù)測(cè)系統(tǒng)的架構(gòu)，可以用于大路網(wǎng)、多點(diǎn)位實(shí)時(shí)預(yù)測(cè)。實(shí)際數(shù)據(jù)實(shí)驗(yàn)和仿真實(shí)驗(yàn)都表明本文提出的預(yù)測(cè)系統(tǒng)能夠應(yīng)對(duì)大路網(wǎng)、高數(shù)據(jù)流量條件下的預(yù)測(cè)，同時(shí)能取得較好的預(yù)測(cè)效果。 二：高速公路事故時(shí)空分布規(guī)律分析及事故預(yù)警算法研究。以津薊高速的事故數(shù)據(jù)為基礎(chǔ)，對(duì)高速公路交通事故時(shí)空分布規(guī)律進(jìn)行了分析，發(fā)現(xiàn)高速公路上的交通事故與一天中的時(shí)段、日期、地點(diǎn)、車(chē)型比例和交通流量都有明顯的關(guān)系，證明了高速公路事故的時(shí)空規(guī)律性。提出了一種基于模糊推理的高速公路交通事故的預(yù)警算法，通過(guò)對(duì)實(shí)時(shí)采集的交通流參數(shù)進(jìn)行分析，識(shí)別交通流的異常狀態(tài)，基于天津市實(shí)測(cè)數(shù)據(jù)的實(shí)驗(yàn)表明，該方法能夠比較準(zhǔn)確的預(yù)測(cè)交通事故風(fēng)險(xiǎn)，適合作為管理者決策參考。 三：高速公路事故預(yù)先應(yīng)對(duì)方法。提出了一種高速公路巡邏車(chē)輛的優(yōu)化調(diào)度算法。以車(chē)輛路徑規(guī)劃問(wèn)題為基礎(chǔ)，將預(yù)測(cè)的路段事故風(fēng)險(xiǎn)加入目標(biāo)函數(shù)中，隨后設(shè)計(jì)了一種兩階段的啟發(fā)式算法進(jìn)行求解，并在不同高速公路網(wǎng)絡(luò)狀態(tài)條件下進(jìn)行了實(shí)驗(yàn)。證明該算法能夠提高巡邏車(chē)輛對(duì)事故路段的覆蓋率，提高事故反應(yīng)時(shí)間。 綜上所述，本文針對(duì)高速公路交通事故管理中的“事前處理”階段展開(kāi)研究，提出了涵蓋交通流預(yù)測(cè)、事故預(yù)警、預(yù)警之后的應(yīng)對(duì)處理等階段的若干方法�；趯�(shí)際數(shù)據(jù)和仿真軟件的實(shí)驗(yàn)表明，，本文提出的方法對(duì)特定問(wèn)題能取得較好的效果，具有一定的可行性。
[Abstract]:It is of great significance to improve the management level of expressway accidents to ensure the healthy development of national economy, the safety and comfort of people's travel, and the peace and harmony of social life. At first, this paper analyzes the stages of traffic accident management from the angle of "crisis management", puts forward the frame structure of expressway traffic accident management system, and summarizes the functions of each part of the frame. This paper summarizes the current research situation related to each basic functional part, and selects the "accident prior treatment" section, which is relatively weak in current research, as the focal point of this paper. According to the logical level of traffic flow forecasting, traffic accident warning and accident early response, the main research contents and results are as follows: first, traffic flow short-term forecasting method and forecasting system research, Firstly, a state vector selection algorithm based on cohesion index and genetic algorithm is proposed, and the effect of the algorithm is verified in four different scenarios. It is proved that the proposed algorithm can significantly improve the accuracy of nonparametric regression prediction and neural network prediction. Then, a data driven short time traffic flow prediction system is proposed, which can be used for real time prediction of large road networks and multi points. The actual data experiments and simulation experiments show that the proposed forecasting system can deal with the prediction under the condition of large road network and high data flow, and can achieve good prediction results. Second, the analysis of temporal and spatial distribution of expressway accidents and the research of accident warning algorithm. Based on the accident data of artichoke expressway, the temporal and spatial distribution of expressway traffic accidents is analyzed. It is found that the traffic accidents on the freeway are related to the period, date and place of the day. The proportion of vehicle type and traffic flow are obviously related, which proves the temporal and spatial regularity of highway accidents. This paper presents an early warning algorithm for freeway traffic accidents based on fuzzy reasoning. By analyzing the traffic flow parameters collected in real time, the abnormal state of traffic flow is identified. The experimental results based on the measured data in Tianjin show that, This method can accurately predict the risk of traffic accidents and is suitable for decision-making of managers. Three: highway accident in advance. This paper presents an optimal scheduling algorithm for highway patrol vehicles. Based on the vehicle path planning problem, the predicted road accident risk is added to the objective function, and then a two-stage heuristic algorithm is designed to solve the problem, and experiments are carried out under different expressway network conditions. It is proved that the algorithm can improve the coverage of patrol vehicles and improve the response time. To sum up, this paper studies the stage of "prior treatment" in expressway traffic accident management, and puts forward some methods including traffic flow prediction, early warning of accidents, and response to early warning. Experiments based on actual data and simulation software show that the method proposed in this paper can achieve good results for specific problems and has certain feasibility.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2014
【分類(lèi)號(hào)】：U491

【參考文獻(xiàn)】

相關(guān)期刊論文 前7條

1 劉榮;;我國(guó)高速公路管理體制存在的問(wèn)題及對(duì)策研究[J];產(chǎn)業(yè)與科技論壇;2010年01期

2 李相勇,張南,蔣葛夫;道路交通事故灰色馬爾可夫預(yù)測(cè)模型[J];公路交通科技;2003年04期

3 陳鐵明;龔榮盛;SAMUEL H Huang;;一種新的快速模糊規(guī)則提取方法[J];控制與決策;2008年09期

4 孫才志,王敬東,潘俊;模糊聚類(lèi)分析最佳聚類(lèi)數(shù)的確定方法研究[J];模糊系統(tǒng)與數(shù)學(xué);2001年01期

5 鄭黎黎;丁同強(qiáng);范海燕;楊兆升;孫健;;高速公路交通事件影響范圍的模糊預(yù)測(cè)[J];數(shù)學(xué)的實(shí)踐與認(rèn)識(shí);2009年01期

6 姜紫峰，韓錫令;高速公路交通流中等密度區(qū)的可變速度和入口匝道聯(lián)合控制[J];西安公路交通大學(xué)學(xué)報(bào);1994年04期

7 馬壽峰,賀國(guó)光,劉豹;智能交通系統(tǒng)中短時(shí)交通流預(yù)測(cè)系統(tǒng)的研究[J];預(yù)測(cè);2004年02期

本文編號(hào)：2123979