臺北市非號誌化路口肇事特性分析與改善策略研擬

Analysis of Traffic Accident and Safety Improvement Measures for Unsignalized Intersection in Taipei City

謝易達(Yi-Da Xie)

指導教授 ： 黃承傳

國立交通大學/管理學院/交通運輸研究所/碩士(2010年)

https://doi.org/10.6842/NCTU.2010.00432

摘要

交通事故地點可區分為「路段」及「路口」兩大類，路口又因交通管制方式之不同，而可區分為號誌化路口與非號誌化路口；在主要道路匯集的交岔路口，因交通量較大，必須設置號誌分配時相，供不同方向之車流通行，但在交通量未達設置標準之路口，則以設置警告或禁制標誌（線）等設施，提升路口的顯著性或區分車輛路權等，以避免交通事故發生。鑑於非號誌化路口與號誌化路口之態樣不盡相同，但以往研究常將兩者合併，無法區別號誌化與非號誌化路口之交通事故特性，故本研究特別針對非號誌路口之交通事故為深入探討，找出影響肇事嚴重程度之因素，做為研擬提升非號誌化路口交通安全的基本依據。
本研究以民國97年臺北市非號誌化路口交通事故為分析對象，將肇事嚴重程度分為「死亡或骨折以上」、「人員受傷」及「單純車輛損壞」3類，並依文獻回顧及實務上所能取得的資料，蒐集「人」、「車」、「路」、「環境」及「交通管制因素」等19項研究變數；先構建具有最多顯著變數之多項羅吉特模式，再以巢式羅吉特模式校估嚴重程度之相似性，及透過混合羅吉特模式校估參數之異質性後，做為研擬改善策略之參考。
依混合羅吉特模式校估結果，對於肇事後致人死亡或骨折以上之案件，以「同向車流之交叉衝突」等6個變數有顯著正向影響，另以「網狀線」等5個變數有顯著負向影響，且均為固定參數；而對於肇事致人員受傷之案件，以「不同方向車流之交叉衝突」等9個變數呈顯著正向影響，其中「不同方向車流之交叉衝突」、「相同方向車流之交叉衝突」及「閃光號誌」為隨機參數，表示這3個變數在「人員受傷」部分具有異質性；而「第一當事人性別」等5個變數則呈顯著負向影響。在改善策略研擬部分，交通工程建議從提升路口能見度、改變路口10公尺內禁止臨停標線及於交通標誌引進ITS技術著手，並選擇地點試辦，檢視其設置成效；交通執法則建議應強化閃光號誌及停讓設施的管制力、取締路口10公尺內停車等，以建立正確的用路權念；在教育宣導則建議可宣導駕駛人多加利用路口反射鏡、轉彎前應注意左（右）後方來車等，以提升駕駛人交通安全觀念。

關鍵字

非號誌化路口 ； 混合羅吉特模式 ； 肇事嚴重程度

並列摘要

The accidents could be sorted to section and intersection by locations, and the intersections could be classified into signalized and unsignalized intersections. Intersections with high volume of traffic usually use traffic singals to separate the vehicles from different direction. But the volume of traffic in intersection without exceeding the criteria usually only install warning or prohibition sign (or marking) to promote visibility of intersection and confirm the right of way. In consideration of the accidents in signalized or unsignalized are different, this study focuses on the accidents of unsignalized intersections to find the influential factors of accident severity, and to propose the safety improvement measures.
This study is based the accident data of unsignalized intersection in 2008 in Taipei City. The severity is classified into three categories: “property damage only”, “injury” and “incapacitating and fatal injury". Depending on literature review and practical experience, there are 19 possible influential factors including “driver”, “vehicle”, “road”, “environment” and “traffic control”. Through the development of a most suitable logit model with maximum number of factors, we then explore the similarity of the classification of severity by nest logit model, and to investigate the difference of parameters by mixed logit model.
The results show that mixed logit model can well describe the accident severity for unsignalized intersections. For incapacitating and fatal injury, there are six positive variables (ex. cross between the same traffic flow etc.) and five negative variables (ex. net marking etc.). For injury, there are nine positive variables (ex. cross between the differ traffic flow etc.) and five negative variables (ex. male etc.).In this study, the proposed safety improvement measures in traffic engineering are to promote visibility of intersection, change the marking for parking near the intersection and integrate ITS with traffic sign. In enforcement, the measures are to intensify the effectiveness for flashing light and stop or yield sign (marking), ban parking near the intersection. Meanwhile, the measures in education are to promote the habit of drivers to use the reflectors and look out the rear vehicles before turning etc.

並列關鍵字

Unsignalized intersection ； Mixed logit model ； Accident severity

參考文獻

23.Jinn-Tsai Wong, Yi-Shih Chung, Shih-Hsuan Huang（2010）,“Determinants behind young motorcyclists’ risky riding behavior”,Accident Analysis and Prevention, Vol.42, Issue 1, pp.275-281.

25.Al-Ghamdi, Ali S.（2002）,“Using logistic regression to estimate the influence of accident”,Accident Analysis and Prevention, Vol.34, No.6, pp.729-741.

26.Mussone, L., Ferrari, A. and Oneta, M., (1999) “An analysis of urban collisions using an artificial intelligence model”, Accident Analysis and Prevention, Vol. 31, pp.705-718.

28.Levine, N., Kim, K. E., & Nitz, L. H. (1995). Daily fluctuations in Honolulu motor vehicle accidents. Accident Analysis and Prevention, Vol.27, No.6, pp.785-796.

29.Joshua, S., & Garber, N. (1990). Estimating truck accident rate and involvement using linear and Poisson regression models. Transportation Planning and Technology, Vol.15, pp.41-58.