基于車檢器及收費數(shù)據(jù)融合的高速公路異常狀態(tài)識別研究
發(fā)布時間：2018-07-15 10:53
【摘要】：異常狀態(tài)識別是高速公路運管部門進行運營管控、路況信息發(fā)布和交通誘導的基礎，對減少交通事故造成的人身傷亡、財產(chǎn)損失和避免二次交通事故等方面具有重要的作用。隨著智能交通技術的發(fā)展和交通信息檢測手段的增加，如何利用交通信息檢測手段進行異常狀態(tài)自動識別引起了廣泛關注。因此，研究高速公路異常狀態(tài)自動識別的關鍵技術對提升高速公路管理水平和服務水平具有重要的理論和實際意義。 本文重點針對目前高速公路車輛檢測器布設數(shù)量不足導致算法實際應用效果較差的問題，從現(xiàn)有高速交通信息檢測手段出發(fā)，分析高速公路收費數(shù)據(jù)，提出了基于車檢器和收費數(shù)據(jù)融合的異常狀態(tài)識別方法。首先重點分析了基于車檢器的交通狀態(tài)參數(shù)對異常狀態(tài)的靈敏度，同時考慮多種數(shù)據(jù)源進行異常狀態(tài)識別，采用收費數(shù)據(jù)建立異常狀態(tài)識別方法。最后利用信息融合，建立基于車檢器及收費數(shù)據(jù)融合的異常狀態(tài)識別方法，解決了單一數(shù)據(jù)源異常狀態(tài)識別效果較差的問題。論文的主要研究工作如下： ①交通狀態(tài)參數(shù)對異常狀態(tài)的靈敏度分析。本文對異常狀態(tài)下交通狀態(tài)參數(shù)進行靈敏度分析，為基于車檢器數(shù)據(jù)的ACI算法參數(shù)選擇奠定基礎。靈敏度分析主要考慮了流量和異常交通狀態(tài)嚴重程度兩個因素的影響，通過仿真表明，交通狀態(tài)參數(shù)可靈敏的反映交通狀態(tài)的變化。 ②針對現(xiàn)有高速公路車輛檢測器布設數(shù)量不足，本文從多源交通信息采集方式出發(fā)，根據(jù)高速公路收費數(shù)據(jù)特征，建立基于收費數(shù)據(jù)的異常狀態(tài)識別方法。由于算法性能受到樣本車輛數(shù)影響，因此本文對算法進行改進，并選取實際數(shù)據(jù)進行驗證，結果表明改進算法在低流量情況下的狀態(tài)識別性能有所提高。 ③針對基于單一數(shù)據(jù)源的異常狀態(tài)識別可能存在可信度低、實際應用效果較差等問題，本文考慮了基于不同數(shù)據(jù)源融合進行異常狀態(tài)識別。本文將算法表決融合方法引入異常狀態(tài)識別領域，建立基于算法表決融合的異常狀態(tài)識別方法，，算法主要包括3部分：基于車檢器數(shù)據(jù)的ACI算法模塊、基于收費數(shù)據(jù)的ACI算法模塊和算法表決融合模塊。最后，本文采用實際數(shù)據(jù)進行驗證，結果表明，相比基于單一數(shù)據(jù)源的狀態(tài)識別算法，融合算法的在異常狀態(tài)識別精度方面具有較好的性能。
[Abstract]:The recognition of abnormal state is the basis of highway transportation and management department for operation control, road condition information release and traffic guidance. It plays an important role in reducing casualties caused by traffic accidents, property losses and avoiding secondary traffic accidents. With the development of intelligent transportation technology and the increase of traffic information detection methods, how to use traffic information detection means to automatically identify abnormal state has attracted much attention. Therefore, it is of great theoretical and practical significance to study the key technology of automatic identification of abnormal state of expressway to improve the level of expressway management and service. In this paper, aiming at the problem that the number of vehicle detectors in freeway is not enough and the practical application effect of the algorithm is poor, this paper analyzes the toll data of freeway from the existing means of high-speed traffic information detection. An abnormal state recognition method based on vehicle detector and toll data fusion is proposed. Firstly, the sensitivity of traffic state parameters based on vehicle detector to abnormal state is analyzed. At the same time, several data sources are considered to identify abnormal state, and the method of abnormal state identification is established by using toll data. Finally, using information fusion, a method of abnormal state recognition based on vehicle detector and toll data fusion is established, which solves the problem of poor recognition effect of single data source. The main work of this paper is as follows: 1 sensitivity analysis of traffic state parameters to abnormal state. In this paper, the sensitivity analysis of traffic state parameters in abnormal state is carried out, which lays a foundation for the parameter selection of ACI algorithm based on vehicle detector data. The sensitivity analysis mainly considers the influence of two factors, the volume of traffic and the severity of abnormal traffic state, and the simulation results show that, Traffic state parameters can reflect the change of traffic state sensitively. (2) in view of the insufficient number of existing highway vehicle detectors, this paper starts from the multi-source traffic information collection method, according to the characteristics of highway toll data. A method of identifying abnormal state based on toll data is established. Because the performance of the algorithm is affected by the sample number of vehicles, this paper improves the algorithm, and selects the actual data to verify the algorithm. The results show that the improved algorithm can improve the performance of state recognition in the case of low traffic. 3 in view of the problems of low reliability and poor practical application of abnormal state recognition based on single data source, In this paper, anomaly state identification based on fusion of different data sources is considered. In this paper, the algorithm of voting fusion is introduced into the field of abnormal state recognition, and the algorithm of abnormal state recognition based on algorithm voting fusion is established. The algorithm mainly includes three parts: ACI algorithm module based on vehicle detector data. ACI algorithm module and algorithm voting fusion module based on charging data. Finally, the actual data is used to verify the proposed algorithm. The results show that the fusion algorithm has better performance than the state recognition algorithm based on a single data source in terms of the accuracy of abnormal state recognition.
【學位授予單位】：重慶大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：U495;U491.116
本文編號：2123828
[Abstract]:The recognition of abnormal state is the basis of highway transportation and management department for operation control, road condition information release and traffic guidance. It plays an important role in reducing casualties caused by traffic accidents, property losses and avoiding secondary traffic accidents. With the development of intelligent transportation technology and the increase of traffic information detection methods, how to use traffic information detection means to automatically identify abnormal state has attracted much attention. Therefore, it is of great theoretical and practical significance to study the key technology of automatic identification of abnormal state of expressway to improve the level of expressway management and service. In this paper, aiming at the problem that the number of vehicle detectors in freeway is not enough and the practical application effect of the algorithm is poor, this paper analyzes the toll data of freeway from the existing means of high-speed traffic information detection. An abnormal state recognition method based on vehicle detector and toll data fusion is proposed. Firstly, the sensitivity of traffic state parameters based on vehicle detector to abnormal state is analyzed. At the same time, several data sources are considered to identify abnormal state, and the method of abnormal state identification is established by using toll data. Finally, using information fusion, a method of abnormal state recognition based on vehicle detector and toll data fusion is established, which solves the problem of poor recognition effect of single data source. The main work of this paper is as follows: 1 sensitivity analysis of traffic state parameters to abnormal state. In this paper, the sensitivity analysis of traffic state parameters in abnormal state is carried out, which lays a foundation for the parameter selection of ACI algorithm based on vehicle detector data. The sensitivity analysis mainly considers the influence of two factors, the volume of traffic and the severity of abnormal traffic state, and the simulation results show that, Traffic state parameters can reflect the change of traffic state sensitively. (2) in view of the insufficient number of existing highway vehicle detectors, this paper starts from the multi-source traffic information collection method, according to the characteristics of highway toll data. A method of identifying abnormal state based on toll data is established. Because the performance of the algorithm is affected by the sample number of vehicles, this paper improves the algorithm, and selects the actual data to verify the algorithm. The results show that the improved algorithm can improve the performance of state recognition in the case of low traffic. 3 in view of the problems of low reliability and poor practical application of abnormal state recognition based on single data source, In this paper, anomaly state identification based on fusion of different data sources is considered. In this paper, the algorithm of voting fusion is introduced into the field of abnormal state recognition, and the algorithm of abnormal state recognition based on algorithm voting fusion is established. The algorithm mainly includes three parts: ACI algorithm module based on vehicle detector data. ACI algorithm module and algorithm voting fusion module based on charging data. Finally, the actual data is used to verify the proposed algorithm. The results show that the fusion algorithm has better performance than the state recognition algorithm based on a single data source in terms of the accuracy of abnormal state recognition.
【學位授予單位】：重慶大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：U495;U491.116
【參考文獻】
相關期刊論文 前10條
1 崔平遠,黃曉瑞;基于聯(lián)合卡爾曼濾波的多傳感器信息融合算法及其應用[J];電機與控制學報;2001年03期
2 張雷元;袁建華;趙永進;;道路交通狀態(tài)識別技術研究[J];道路交通與安全;2009年02期
3 曲陽;徐林;王建輝;顧樹生;;基于信息博弈的多源信息融合方法[J];東南大學學報(自然科學版);2005年S2期
4 王力,王川久,沈曉蓉,范躍祖;智能交通系統(tǒng)中實時交通信息采集處理的新方法[J];系統(tǒng)工程;2005年02期
5 李雁;陸海亭;張寧;;一種短時交通流異常數(shù)據(jù)識別新算法[J];公路交通科技(應用技術版);2010年04期
6 伍建國,王峰;城市道路交通數(shù)據(jù)采集系統(tǒng)檢測器優(yōu)化布點研究[J];公路交通科技;2004年02期
7 孫棣華;陳偉霞;;基于手機定位的高速公路事件檢測方法研究[J];公路交通科技;2006年02期
8 馬驥;裴玉龍;;智能交通系統(tǒng)(ITS)信息采集技術評述[J];哈爾濱工業(yè)大學學報;2003年01期
9 孫即祥,史慧敏,王宏強;信息融合中的有關熵理論[J];計算機學報;2003年07期
10 熊文華;徐建閩;林思;;基于BP網(wǎng)絡的浮動車與線圈檢測數(shù)據(jù)融合模型[J];計算機仿真;2009年09期
本文編號：2123828
本文鏈接：http://sikaile.net/kejilunwen/jiaotonggongchenglunwen/2123828.html
