摘要
基于机非冲突技术,根据冲突数量和严重程度采用模糊聚类法将甘家口交叉口冲突区分成了3类;选取冲突级别较为严重的冲突点—直行自行车与相邻车道的右转机动车的冲突为研究对象,从自行车和右转机动车的运行特性、违章行驶行为、交通环境等方面深入地分析交叉口引起冲突的安全因素;从安全工程学的角度,以交叉口交通冲突为事故树的起点建立平面交叉口机非冲突事故树,并对其进行定性定量分析;根据事故树分析结果,把事故树的最小割集作为输入样本建立平面交叉口机非冲突模糊神经网络模型,用Matlab编程实现了对冲突和引起冲突的各原因间关系的预测,最后进行实例验证表明该模型可以较好地分析、预测和评价引起平面交叉口机-非冲突的主要原因。
The paper classifies the at-grade intersections into three groups in terms of conflict occurrence and seriousness using fuzzy clustering method. In an empirical study, it applies the procedure to the at-grade intersections of Ganjiakou district of Beijing and then chooses the intersections with the most serious conflicts as the study object of the paper. It first analyzes the safety hazards that could lead to conflicts and then studies quantitatively the motor-bicycle conflict failure tree of the intersections, on the basis of which a fuzzy neural network model for the conflicts is set up. The model can identify the causes of a conflict as it happens and at the end, an empiriealy study is carried out to verify the validity of the model.
出处
《物流技术》
2011年第1期56-59,共4页
Logistics Technology
关键词
城市交通
机-非交通冲突
模糊神经网络
事故树
平面交叉口
urban traffic
motor-bicycle conflict
fuzzy neural network
failure tree
at-grade intersection
