基于交通沖突技術(shù)的無信號控制交叉口安全風(fēng)險自動評價
發(fā)布時間：2018-08-05 16:05
【摘要】：交叉口是道路網(wǎng)的重要組成部分,也是交通事故頻發(fā)的地點。為提高交叉口的安全管理水平,提出合理的改善方案、比較改善前后的效果,首先需對其進行交通安全評價。相對傳統(tǒng)的事故評價法而言,基于交通沖突技術(shù)(Traffic Conflict Techniques,TCT)的評價方法具有“大樣本、短周期、高信度”的特點,因而備受到研究者青睞。目前我國基于交通沖突技術(shù)的交義口安全評價方法中,定性分析多、定量分析較少；人工觀測多、沖突的自動識別和采集較少。為此,本文以交通沖突的自動判別和記錄為應(yīng)用背景,建立了無信號控制交叉口危險度評價模型。 借鑒國內(nèi)外已有的研究成果,論文首先系統(tǒng)的闡述了交通沖突和交通事故之間的關(guān)系,探討了采用交通沖突評價交叉口安全的可行性與合理性。通過分析沖突產(chǎn)生的機理,給出了適用于機器自動識別場合的交通沖突的定義；分析了影響交通沖突嚴重程度的因素,給出了評價沖突嚴重程度的思路。 其次,論文利用最長公共子序列算法在處理不同長度的軌跡以及軌跡異值點上的優(yōu)越性,將傳統(tǒng)的最長公共子序列算法由名義序列的相似程度比較拓展至車輛軌跡的相似程度的比較,從而實現(xiàn)典型軌跡模式的提取。提出綜合利用碰撞時間和車輛速度的“動態(tài)閾值法”判斷兩條預(yù)期軌跡是否為碰撞軌跡,根據(jù)兩車軌跡的類型完成對潛在沖突的判斷和識別。以車輛預(yù)測軌跡的條件概率和相應(yīng)軌跡對應(yīng)的碰撞時間為自變量,描述交通沖突與交通事故的接近程度；用相互作用的兩車速度描述潛在事故的嚴重程度,建立了衡量交通沖突嚴重程度的評價模型。將其推廣應(yīng)用至整個交叉口,建立了評價整個交叉口的安全水平的自動分析模型。 最后,采用微觀交通流仿真軟件VISSIM和數(shù)學(xué)分析軟件Matlab作為仿真工具,以某T型無信號控制交叉口為例,對其實施安全改善措施前、后的安全風(fēng)險水平進行了比較,對改善措施的有效性進行了評價。研究結(jié)果表明該模型可以應(yīng)用于平面無信號控制交叉口的交通安全自動評價,對于改善交叉口的安全狀況具有重要的指導(dǎo)意義。
[Abstract]:Intersection is an important part of road network and also a place where traffic accidents occur frequently. In order to improve the safety management level of intersections, a reasonable improvement scheme is put forward, and the effect before and after improvement is compared. The first step is to evaluate the traffic safety of intersections. Compared with the traditional accident evaluation method, the evaluation method based on traffic conflict technology (Traffic Conflict technique TCT has the characteristics of "large sample, short period, high confidence", so it is favored by researchers. At present, there are more qualitative analysis, less quantitative analysis, more manual observation and less automatic identification and collection of conflicts in the safety evaluation methods based on traffic conflict technology. Therefore, based on the application background of automatic discrimination and recording of traffic conflicts, a risk assessment model for signal-free intersections is established in this paper. Referring to the existing research results at home and abroad, this paper first systematically expounds the relationship between traffic conflicts and traffic accidents, and probes into the feasibility and rationality of using traffic conflicts to evaluate the safety of intersections. By analyzing the mechanism of conflict, the definition of traffic conflict suitable for machine automatic recognition is given, the factors influencing the severity of traffic conflict are analyzed, and the train of thought for evaluating the severity of conflict is given. Secondly, the paper uses the longest common subsequence algorithm to deal with the different length of trajectory and locus outliers. The traditional longest common subsequence algorithm is extended from the similarity comparison of nominal sequences to the comparison of the similarity degree of vehicle trajectories so as to achieve the extraction of typical trajectory patterns. A dynamic threshold method based on collision time and vehicle velocity is proposed to judge whether the two expected tracks are collision tracks, and to judge and identify the potential conflicts according to the type of the two vehicle tracks. The conditional probability of vehicle trajectory prediction and the collision time corresponding to the corresponding trajectory are taken as independent variables to describe the proximity between traffic conflicts and traffic accidents, and the severity of potential accidents is described by the two vehicle velocities that interact with each other. An evaluation model is established to measure the severity of traffic conflict. The automatic analysis model for evaluating the safety level of the whole intersection is established by applying it to the whole intersection. Finally, the microcosmic traffic flow simulation software VISSIM and the mathematical analysis software Matlab are used as simulation tools. Taking a T type unsignalized intersection as an example, the safety risk levels before and after the implementation of safety improvement measures are compared. The effectiveness of the improvement measures is evaluated. The results show that the model can be applied to the automatic evaluation of traffic safety at planar unsignalized intersections, and it is of great significance to improve the safety of intersections.
【學(xué)位授予單位】：南京理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：U491
本文編號：2166295
[Abstract]:Intersection is an important part of road network and also a place where traffic accidents occur frequently. In order to improve the safety management level of intersections, a reasonable improvement scheme is put forward, and the effect before and after improvement is compared. The first step is to evaluate the traffic safety of intersections. Compared with the traditional accident evaluation method, the evaluation method based on traffic conflict technology (Traffic Conflict technique TCT has the characteristics of "large sample, short period, high confidence", so it is favored by researchers. At present, there are more qualitative analysis, less quantitative analysis, more manual observation and less automatic identification and collection of conflicts in the safety evaluation methods based on traffic conflict technology. Therefore, based on the application background of automatic discrimination and recording of traffic conflicts, a risk assessment model for signal-free intersections is established in this paper. Referring to the existing research results at home and abroad, this paper first systematically expounds the relationship between traffic conflicts and traffic accidents, and probes into the feasibility and rationality of using traffic conflicts to evaluate the safety of intersections. By analyzing the mechanism of conflict, the definition of traffic conflict suitable for machine automatic recognition is given, the factors influencing the severity of traffic conflict are analyzed, and the train of thought for evaluating the severity of conflict is given. Secondly, the paper uses the longest common subsequence algorithm to deal with the different length of trajectory and locus outliers. The traditional longest common subsequence algorithm is extended from the similarity comparison of nominal sequences to the comparison of the similarity degree of vehicle trajectories so as to achieve the extraction of typical trajectory patterns. A dynamic threshold method based on collision time and vehicle velocity is proposed to judge whether the two expected tracks are collision tracks, and to judge and identify the potential conflicts according to the type of the two vehicle tracks. The conditional probability of vehicle trajectory prediction and the collision time corresponding to the corresponding trajectory are taken as independent variables to describe the proximity between traffic conflicts and traffic accidents, and the severity of potential accidents is described by the two vehicle velocities that interact with each other. An evaluation model is established to measure the severity of traffic conflict. The automatic analysis model for evaluating the safety level of the whole intersection is established by applying it to the whole intersection. Finally, the microcosmic traffic flow simulation software VISSIM and the mathematical analysis software Matlab are used as simulation tools. Taking a T type unsignalized intersection as an example, the safety risk levels before and after the implementation of safety improvement measures are compared. The effectiveness of the improvement measures is evaluated. The results show that the model can be applied to the automatic evaluation of traffic safety at planar unsignalized intersections, and it is of great significance to improve the safety of intersections.
【學(xué)位授予單位】：南京理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：U491
【參考文獻】
相關(guān)期刊論文 前10條
1 成衛(wèi),丁同強,李江;道路交叉口交通沖突灰色評價研究[J];公路交通科技;2004年06期
2 葉凡,陸鍵,丁紀平,項喬君;交通沖突技術(shù)在ETC安全評價中的應(yīng)用研究[J];公路交通科技;2004年12期
3 劉小明,段海林;平面交叉口交通沖突技術(shù)標準化研究[J];公路交通科技;1997年03期
4 裴玉龍;馮樹民;;基于交通沖突的行人過街危險度研究[J];哈爾濱工業(yè)大學(xué)學(xué)報;2007年02期
5 曲昭偉;李志慧;胡宏宇;郭偉偉;魏巍;;基于視頻處理的無信號交叉口交通沖突自動判別方法[J];吉林大學(xué)學(xué)報(工學(xué)版);2009年S2期
6 張學(xué)亮;鄧衛(wèi);郭唐儀;;基于沖突率的交叉口交通安全評價方法研究[J];交通運輸工程與信息學(xué)報;2007年01期
7 趙永紅;白玉;楊曉光;;基于Poisson過程的交通沖突預(yù)測模型研究[J];交通信息與安全;2011年01期
8 劉淼淼;魯光泉;王云鵬;田大新;;交叉口交通沖突嚴重程度量化方法[J];交通運輸工程學(xué)報;2012年03期
9 高新中;洪成杰;;高速公路養(yǎng)護作業(yè)交通沖突模型分析[J];黑龍江科技信息;2013年07期
10 汪瑩;黃新;;交叉口交通安全評價方法回顧與分析[J];森林工程;2013年02期
，本文編號：2166295
本文鏈接：http://www.sikaile.net/kejilunwen/jiaotonggongchenglunwen/2166295.html
教材專著
