基于模糊逻辑的信号交叉口人车冲突判别方法

Fuzzy logic based pedestrian-vehicle conflict discrimination at signalized intersection

利用视频图像处理技术提取交叉口人车冲突基本数据。采用冲突时间差和车辆速度两个特征对人车冲突安全性进行分析,并由此构建相应的隶属度函数,确定了人车冲突安全等级,设计了基于模糊逻辑规则的信号交叉口人车冲突判别方法。通过与人工判别结果对比,验证了本文方法的有效性,本文建立的基于模糊逻辑的人车冲突判别模型准确率超过85%。

Image processing technology was used to extract the basic data of pedestrian-vehicle conflict. Time Difference to Collision （TDTC） and the vehicle speed were selected as the features for pedestrian-vehicle conflict analysis. Based on the analysis, the corresponding membership functions of the two features are constructed to distinguish the safety level of the conflict. The proposed method is compared with of artificial discrimination. Results demonstrate that the proposed method is reliable and has good performance. The discriminating accuracy of the method is higher than 85 %.