基于OBD數(shù)據(jù)分析的駕駛行為研究
發(fā)布時間：2019-03-08 20:21
【摘要】：近年來,隨著社會經(jīng)濟發(fā)展,國內(nèi)汽車保有量也在大幅度增加,加強交通安全管理從而有效防止事故的發(fā)生成為近年來交通問題研究的重點。駕駛員作為車輛的操作者,其駕駛行為是交通運輸安全影響的最主要因素,研究駕駛行為對預防交通事故、促進智能交通及駕駛輔助系統(tǒng)的發(fā)展具有重要意義。本文運用車聯(lián)網(wǎng)相關技術,通過OBD智能車載終端獲取車輛行駛數(shù)據(jù),采用算法對數(shù)據(jù)進行挖掘分析,辨識駕駛員的駕駛行為和車輛狀態(tài),對異常駕駛行為進行算法識別,依據(jù)駕駛行為指標及其他行程數(shù)據(jù)對駕駛傾向性進行識別和構建駕駛行為評分模型,文章主要內(nèi)容如下:(1)介紹了駕駛行為數(shù)據(jù)的獲取和簡單的預處理過程。對文中采集數(shù)據(jù)所用的OBD設備車載終端構造、安裝位置等做了簡單介紹,詳細介紹了設備中的數(shù)據(jù)采集模塊,并對相關數(shù)據(jù)做了預處理。(2)在駕駛行為評價指標研究的基礎上,提出了超速、急加速、急減速、急剎車、急轉(zhuǎn)彎、高轉(zhuǎn)速、轉(zhuǎn)速不匹配、長怠速等異常駕駛行為識別原理算法及判斷流程,通過實車實驗數(shù)據(jù)分析驗證駕駛行為算法識別效果,并對超速行為、急變速行為做詳細識別結(jié)果分析。(3)對駕駛傾向性相關理論進行闡述,提取駕駛傾向性識別所需要的車輛狀態(tài)數(shù)據(jù)和異常駕駛行為數(shù)據(jù)等特征參數(shù),先利用k-means聚類算法對駕駛傾向性進行分類,再使用BP神經(jīng)網(wǎng)絡對駕駛傾向性樣本進行訓練及評估,并驗證評估效果的可靠性。(4)構建駕駛行為評分模型,使用改進型熵權層次分析法AEW-AHP確定指標權重,相比單一的熵權法和層次分析法計算結(jié)果表明,AEW-AHP法更加符合實際。計算得出權重后,再分析指標的備選項和分值,構建駕駛行為評分模型并運用實例進行分析。
[Abstract]:In recent years, with the social and economic development, the number of domestic car ownership is also increasing greatly. Strengthening traffic safety management to effectively prevent the occurrence of accidents has become the focus of research on traffic problems in recent years. As a vehicle operator, the driver's driving behavior is the most important factor affecting traffic safety. It is of great significance to study the driving behavior for preventing traffic accidents, promoting the development of intelligent transportation and driving assistance system. In this paper, we use the related technology of vehicle networking, get the vehicle driving data through OBD intelligent vehicle terminal, use the algorithm to mine and analyze the data, identify the driver's driving behavior and vehicle state, identify the abnormal driving behavior, and carry on the algorithm recognition to the abnormal driving behavior. According to the driving behavior index and other travel data, the driving tendency is recognized and the driving behavior scoring model is constructed. The main contents of this paper are as follows: (1) the acquisition of driving behavior data and the simple preprocessing process are introduced. This paper gives a brief introduction to the structure and installation position of the on-board terminal of the OBD equipment used to collect data, and introduces the data acquisition module of the equipment in detail. (2) on the basis of the research on the evaluation index of driving behavior, the paper puts forward that overspeed, rapid acceleration, rapid deceleration, sharp brake, sharp turn, high speed and speed mismatch. The principle algorithm and judging process of abnormal driving behavior recognition such as long idle speed are analyzed and verified by the analysis of real vehicle experiment data, and the identification effect of driving behavior algorithm is verified for overspeed behavior. The results are analyzed in detail. (3) the related theory of driving tendency is expounded, and the characteristic parameters such as vehicle state data and abnormal driving behavior data are extracted, which are needed for the identification of driving tendency. Firstly, the k-means clustering algorithm is used to classify the driving tendency, then the BP neural network is used to train and evaluate the driving tendency samples, and the reliability of the evaluation effect is verified. (4) the driving behavior scoring model is constructed. The improved entropy weight analytic hierarchy process (AEW-AHP) is used to determine the index weight. Compared with the single entropy weight method and the analytic hierarchy process (AHP), the results show that the AEW-AHP method is more practical. After the weight is calculated, the reserve options and scores of the indexes are analyzed, and the driving behavior scoring model is constructed and analyzed with an example.
【學位授予單位】：長安大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：U491.25
[Abstract]:In recent years, with the social and economic development, the number of domestic car ownership is also increasing greatly. Strengthening traffic safety management to effectively prevent the occurrence of accidents has become the focus of research on traffic problems in recent years. As a vehicle operator, the driver's driving behavior is the most important factor affecting traffic safety. It is of great significance to study the driving behavior for preventing traffic accidents, promoting the development of intelligent transportation and driving assistance system. In this paper, we use the related technology of vehicle networking, get the vehicle driving data through OBD intelligent vehicle terminal, use the algorithm to mine and analyze the data, identify the driver's driving behavior and vehicle state, identify the abnormal driving behavior, and carry on the algorithm recognition to the abnormal driving behavior. According to the driving behavior index and other travel data, the driving tendency is recognized and the driving behavior scoring model is constructed. The main contents of this paper are as follows: (1) the acquisition of driving behavior data and the simple preprocessing process are introduced. This paper gives a brief introduction to the structure and installation position of the on-board terminal of the OBD equipment used to collect data, and introduces the data acquisition module of the equipment in detail. (2) on the basis of the research on the evaluation index of driving behavior, the paper puts forward that overspeed, rapid acceleration, rapid deceleration, sharp brake, sharp turn, high speed and speed mismatch. The principle algorithm and judging process of abnormal driving behavior recognition such as long idle speed are analyzed and verified by the analysis of real vehicle experiment data, and the identification effect of driving behavior algorithm is verified for overspeed behavior. The results are analyzed in detail. (3) the related theory of driving tendency is expounded, and the characteristic parameters such as vehicle state data and abnormal driving behavior data are extracted, which are needed for the identification of driving tendency. Firstly, the k-means clustering algorithm is used to classify the driving tendency, then the BP neural network is used to train and evaluate the driving tendency samples, and the reliability of the evaluation effect is verified. (4) the driving behavior scoring model is constructed. The improved entropy weight analytic hierarchy process (AEW-AHP) is used to determine the index weight. Compared with the single entropy weight method and the analytic hierarchy process (AHP), the results show that the AEW-AHP method is more practical. After the weight is calculated, the reserve options and scores of the indexes are analyzed, and the driving behavior scoring model is constructed and analyzed with an example.
【學位授予單位】：長安大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：U491.25
【參考文獻】
相關期刊論文 前10條
1 陳傳燦;徐聰;;汽車OBD系統(tǒng)發(fā)展綜述[J];汽車零部件;2016年07期
2 陳曉娟;樓培德;;基于OBD的車載智能終端現(xiàn)狀及其發(fā)展趨勢[J];軟件;2014年10期
3 彭江琴;劉南杰;仲浩;趙海濤;;基于GID的UBI系統(tǒng)研究[J];微型機與應用;2014年22期
4 劉永濤;喬潔;魏朗;鞏建強;張亞岐;;危險駕駛行為辨識算法研究[J];計算機工程與設計;2014年04期
5 肖獻強;任春燕;王其東;;基于隱馬爾可夫模型的駕駛行為預測方法研究[J];中國機械工程;2013年21期
6 嚴新平;張暉;吳超仲;毛U，
本文編號：2437177
本文鏈接：http://sikaile.net/kejilunwen/daoluqiaoliang/2437177.html
