用于城区交叉口闯红灯预测的车辆轨迹特征参数分析  被引量:1

Analysis of Vehicle Trajectory Characteristic Parameters for Red-light-running Prediction in Arterial Intersections

在线阅读下载全文

作  者:王岚君[1] 张瓅玶[2] 王希勤[1] 

机构地区:[1]清华大学电子工程系,北京100084 [2]加州大学伯克利分校,美国94804

出  处:《公路交通科技》2011年第5期108-112,共5页Journal of Highway and Transportation Research and Development

基  金:国家高技术研究发展计划(八六三计划)项目(2006AA11Z113)

摘  要:通过对比闯红灯预测算法的性能,选择出用于区分车辆走、停行为的车辆轨迹特征参数。为实现对交叉口车辆闯红灯行为预测,根据闯红灯车辆为在红灯亮起后通过停止线的车辆这一定义,提出闯红灯预测算法的两个主要功能:估计车辆到达停止线的时刻;预测车辆到达停止线后的走、停行为。研究结果表明:在速度、加速度、行驶时间和车头时距4个物理量中,车辆的速度对闯红灯行为的预测性能最好;若用速度分别与其他3个物理量联合对车辆走、停行为进行区分,性能有所提高,但并不显著;速度-车头时距组合比以往常用的速度-加速度组合进行闯红灯行为预测的性能更好。By comparing the performance of red-light-running(RLR) prediction algorithm,the vehicle trajectory characteristic parameters for distinguishing the go/stop behaviors were selected.In order to predict the RLR at signalized arterial intersections,based on the definition of RLR,which is a vehicle going through the stop bar after red light onset,two main functions of the RLR predicting algorithm were proposed.They are the estimator of the arrival time at the stop bar,and the predictor of the go/stop behavior after arriving the stop bar.Comparing four variables of the subject vehicle: speed,acceleration,travel time and time headway,the results show that(1) the speed is the most useful parameter in distinguishing the go/stop behaviors;(2) when speed is associated with the other three variables respectively in distinguishing the go/stop behaviors,the performances of the RLR prediction are better than using speed only,but the improvement is not very significant;(3) the performance of the RLR prediction based on the vehicle trajectory parameter set with speed and time headway is better than the set with speed and acceleration which were originally used in the existing work.

关 键 词:交通工程 闯红灯预测 统计判决 车辆轨迹 交叉口 

分 类 号:U491[交通运输工程—交通运输规划与管理]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象