透過您的圖書館登入
IP:3.144.113.30
  • 學位論文

應用貝式方法於交叉口左轉決策與風險評估模式

Intersection Left-turn Decision and Risk Assessment Model Using Bayesian Approach

指導教授 : 許添本
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


交叉口左轉管理為交通工程設計重要一環,特別於准許式左轉路口,交叉口左轉車將與對向直行車產生接受間距行為,左轉車駕駛人必須於間距決策時,感知對向來車間程與速度,並評估本車清道時間與對向車到達時間,決定是否通過路口,因此左轉車之間距決策將影響整個交叉口之運行風險。過去臺灣針對接受間距之相關研究,多著重於分析臨界間距與交叉口容量之探討,而未應用於左轉交通安全分析,此外國外研究指出接受間距行為應嚴謹區分為行進接受間距模式與停等接受間距模式,以符合在路口不同之接受間距實際情境,然臺灣目前之相關研究並未深入區分分析。此外接受間距定性分析模式之建立忽略駕駛人之接受間距車流行為不具同質性,而是非定性之隨機模式,亦即不同駕駛人在不同狀況所認知之臨界接受間距並非定值,為克服此缺點,本研究採用具有個體選擇觀念之二元羅吉特模式做為接受間距模式邏輯建構,並分成行進接受間距模式與停等接受間距兩大模式,採用高樓攝影調查分析,以對向來車之車速與空間間程所形成之時間間程及本車等待時間等影響變數,構建左轉車決策接受間距決策邏輯。 本研究分析交叉口左轉汽機車之接受間距決策行為,並進一步探討不同情境下左轉車之決策風險,透過馬可夫鏈蒙地卡羅模擬方法,結合左轉車決策接受間距決策模式與對向車減速煞停模式,評估左轉肇事風險,本研究以肇事機率與肇事嚴重度構建風險指標值,並應用貝式方法於左轉風險評估模式,透過先驗分配資料之分析,經由三階段之馬可夫鏈蒙地卡羅模擬過程,再輸出相關參數之後驗分配資料進行交叉口左轉決策風險評估,以分析汽機車於行進與停等左轉決策之肇事風險,並可實際運用分析至不同速率情境下對左轉肇事風險之影響程度。

並列摘要


Left-turn design plays an important role in traffic engineering, especially in permissive intersection .Gap acceptance behavior will happen when left-turn car driver make gap decision ,based on the spacing gap and speed of opposing car ,and compared the arrival time with clearance time, which influence the operation risk of whole intersection .Most of Taiwan's relevance research in the past focus on the this analysis of critical gap and intersection capacity, not the analysis of left-turn safety. Moreover, according past research, gap acceptance behavior should distinguish lag decision with gap decision, in order to match the reality, which is lacked of analysis in Taiwan's research. On the other hand, deterministic model of gap acceptance ignore the random process of driver decision. Therefore ,this research used logit model as the gap acceptance decision model, which contains lag and gap model. Gap acceptance decision logic is based the relevant variables such as spacing gap and speed through traffic video analysis that influence gap acceptance behavior. The research not only analysis the gap acceptace behavior based on spacing gap and speed in Taiwan but also investigate left turn decision risk in order different scenario. Using the Bayesian approach to combine left-turn car decision and opposing car braking model ,and access the accident risk of left-turn. The research constructs the risk index by using Bayesian approach through three step simulation process, in order to analysis the accident risk of motor and scooter, and access the influence of the speed policy on the left-turn safety of intersection in Taiwan.

參考文獻


46. 孫將瓴,市區道路小汽車變換車道防撞警示系統之研究,國立台灣大學土木工程研究所交通組碩士論文,民國94年。
34. 洪啟峰,混合車流下汽機車車流參數預測模式之研究-渾沌類神經方法之應用,國立台灣大學土木研究所碩士論文,民國93年。
36. 林育瑞,利用類神經網路構建機車車流模式之研究,成功大學交通管理研究所碩士論文,民國90年。
15. Joseph K. Kearney, Timofey Grechkin, James Cremer, Jodie Plamert, Traffic generation for studies of gap acceptance, DSC 2006 Europe Paris October 2006.
38. 曾平毅,張瓊文等,號誌化路口衝突左轉停等車疏解行為之研究,國科會專題報告,民國95年。

被引用紀錄


萬家豪(2014)。混合車流下無號誌圓環容量研究〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2014.02832

延伸閱讀