摘要
为有效提高快速路交通事件检测的覆盖率,解决算法误判率较高的问题,本文通过分析交通事件发生时交通流参数在时间维与空间维的变化,分别提出了基于固定检测器的多参数判别算法与基于浮动车的时空二维判别算法.当两个数据源算法同时满足检测条件时,研究以D-S理论为基础,将两个子算法有效地结合,实现事件的综合检测。最后,研究利用北京市快速路上采集的交通事件数据、固定检测器数据和浮动车数据对算法性能进行了检验。结果表明,基于固定检测器的多参数判别算法和基于浮动车的时空二维判别算法的检测率、误判率都达到了较好的效果,可以满足系统应用的需要。基于D-S理论的综合检测算法,具有比其他经典判别算法与两个子算法更低的算法误判率。
Two detection algorithms based on fixed detector data and floating car data were proposed separately according to the change of traffic parameters in the temporal and spatial dimensions to effectively improve the coverage of expressway automatic incident detection and to reduce false alarm rate （FAR）. Then, the two algorithms were integrated effectively according to the D-S theory to detect the accidents when both algorithms satisfy the detection condition simultaneously. Finally, the algorithms were verified with actual traffic data including the fixed detector data, the floating ear data and the event data collected from Beijing expressway. The results show that both single data source based incident detection algorithms could meet the basic needs of practical application at detection rate （DR） and false alarm rate, and the false alarm rate （FAR） of the integrated detection algorithm based on D-S theory is lower than those of the （FAR） two sub-algorithms and other classical algorithms.
出处
《公路交通科技》
CAS
CSCD
北大核心
2011年第12期112-116,144,共6页
Journal of Highway and Transportation Research and Development
基金
国家“十一五”科技支撑计划项目(2006BAJ18B03)
北京工业大学校青年基金项目
