【摘要】：由于道路交通安全保障形勢嚴峻，行駛車輛對適應(yīng)性強的安全預警系統(tǒng)有著切實的需求。國內(nèi)外在這方面的研究還處于探索性階段，就實用性而言，還有繼續(xù)改進的可能和必要。因此，本文以車輛外流場分布為切入點，綜合運用車輛外流場數(shù)值模擬，行駛軌跡預測和模糊模式識別三種手段，來共同完成對行駛車輛的橫向安全狀態(tài)的預警。安全預警系統(tǒng)的適應(yīng)性和準確性得以提高。 研究所圍繞的主要內(nèi)容包括：首先，對車輛外流場分布進行CFD模擬計算。對影響流場分布的因素對比分析，選定合適的參數(shù)，建立幾個典型模型。分別在直線和曲線行駛下，對單車及多車的外流場進行數(shù)值模擬。驗證線性疊法加獲取車輛外流場分布信息的可行性，，并構(gòu)建數(shù)據(jù)庫。然后，預測行駛車輛的軌跡和獲取車輛狀態(tài)特征。推導車輛的行駛軌跡方程并在MATLAB中求解軌跡曲線。借助道路圖像的灰度特征，擬合出車道邊界線。根據(jù)橫向安全狀態(tài)的各個特征的分析，選出車輛橫向位置、撞線時間及流場各點的縱向流速和橫向流速等四個狀態(tài)特征指標。最后，對車輛安全狀態(tài)模式進行識別。運用模糊動態(tài)聚類對特征樣本集進行模式類別劃分，組建標準的模式類別庫。并據(jù)此，對行駛車輛的實時狀態(tài)特征指標所屬的安全模式類別進行識別和預警。 與傳統(tǒng)車道偏離預警相比，本文提出的車輛安全狀態(tài)識別方法，引入了車輛外流場分布特征，信息指標和適用工況更多，對橫向安全狀態(tài)的判別更精。經(jīng)過檢驗，識別效果比較滿意。
[Abstract]:Because of the serious situation of road traffic safety guarantee, driving vehicles have practical demand for adaptive safety early warning system. The domestic and foreign research in this field is still in the exploratory stage, in terms of practicability, it is possible and necessary to continue to improve. Therefore, this paper takes the vehicle outflow field distribution as the breakthrough point, synthetically uses the vehicle outflow field numerical simulation, the traveling track forecast and the fuzzy pattern recognition three means, completes the traveling vehicle transverse safety state early warning together. The adaptability and accuracy of the security early warning system have been improved. The main contents of the research are as follows: first, the distribution of vehicle outflow field is simulated by CFD. By comparing and analyzing the factors affecting the distribution of flow field, the appropriate parameters are selected and several typical models are established. The flow field of a bicycle and a multi-vehicle is numerically simulated under straight line and curve respectively. To verify the feasibility of linear stacking method to obtain the distribution information of vehicle outflow field, and to construct the database. Then, the trajectory of the vehicle is predicted and the state characteristics of the vehicle are obtained. The vehicle trajectory equation is derived and the trajectory curve is solved in MATLAB. With the help of the grayscale features of the road image, the lane boundary line is fitted. According to the analysis of the characteristics of the transverse safe state, four state characteristic indexes are selected: the transverse position of the vehicle, the time of the collision line and the longitudinal velocity and the transverse velocity of the points in the flow field. Finally, the vehicle safety state pattern is recognized. Fuzzy dynamic clustering is used to classify the feature sample set, and a standard pattern class database is constructed. Based on this, the classification of safety mode which belongs to the real-time state characteristic index of moving vehicle is identified and early warning is carried out. Compared with the traditional lane deviation warning, the vehicle safety state recognition method proposed in this paper introduces the distribution characteristics of the vehicle outflow field, the information index and the applicable working condition, and the discrimination of the lateral safety state is more accurate. After testing, the recognition effect is satisfactory.
【學位授予單位】：燕山大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：U495;TP391.41

【參考文獻】

相關(guān)期刊論文 前10條

1 王榮本;郭烈;金立生;顧柏園;余天洪;;智能車輛安全輔助駕駛技術(shù)研究近況[J];公路交通科技;2007年07期

2 王力;范耀祖;張海;;基于探測車技術(shù)和多級模糊模式識別的道路交通狀態(tài)評價方法[J];公路交通科技;2007年09期

3 孫棣華;王春麗;;基于模糊模式識別的車輛定位地圖匹配算法[J];計算機工程與應(yīng)用;2007年25期

4 武元杰;孟華;;汽車安全車距模型影響因素分析[J];交通與運輸(學術(shù)版);2010年02期

5 王艷瑋,樊其瑾,彭炎午;基于模糊聚類的產(chǎn)品合理子裝配劃分[J];機械科學與技術(shù);2001年02期

6 郭景華;胡平;李琳輝;王榮本;張明恒;郭烈;;基于遺傳優(yōu)化的無人車橫向模糊控制[J];機械工程學報;2012年06期

7 張來平;鄧小剛;張涵信;;動網(wǎng)格生成技術(shù)及非定常計算方法進展綜述[J];力學進展;2010年04期

8 馬雷;武波濤;王連東;;彎曲道路識別方法與目標函數(shù)選取的研究[J];汽車工程;2008年07期

9 馬雷;臧俊杰;張潤生;于福瑩;;基于DS證據(jù)理論的不同光照條件下道路邊界識別[J];汽車工程;2011年08期

10 李克強;王躍建;高鋒;張磊;郭磊;;基于ITS技術(shù)的汽車駕駛安全輔助系統(tǒng)[J];汽車技術(shù);2006年S1期

本文編號：2371858