摘要
为减少交通事件引起的交通延误,有效预防偶发性交通事件导致二次事故的发生,提出一种基于支持向量机(SVM)和数据融合技术的城市道路交通事件自动检测(AID)算法。利用车载激光测距仪采集本车与前车的距离,利用搭载全球定位系统(GPS)的浮动车采集本车瞬时速度。将这2种交通数据按一定的规则进行数据级融合,然后运用线性、多项式和径向基(RBF)3种核函数的SVM模型分别进行事件检测。最后,用实测数据对其进行验证。结果表明:核函数为RBF的非线性SVM模型检测率(DR)值最大,误判率(FAR)值最小,检测指标均优于经典算法,说明算法检测性能良好。
In order to reduce traffic delays caused by traffic incidents and prevent secondary accidents caused by sporadic traffic incidents , a new city roads AID algorithm was worked out based on SVM and data fusion technology. Distance data were collected by using a vehicular laser rangefinder. Instantaneous speed data were collected by using a GPS floating car. The two kinds of data were fused. Then traffic incidents were tested by the SVM models with kernel functions including linear, polynomial and radial basis function (RBF). At last, an illustration was conducted to validate these models by measured data. According to the results of verification, with the nonlinear SVM model with RBF a maximum detection rate(DR) and a mini- mum false alarm rate (FAR) can be obtained and detection indexes can be better than those that can be obtained by using the regular algorithms. It shows that the algorithm has a good detection performance.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2015年第4期81-86,共6页
China Safety Science Journal
基金
深圳市城市规划与决策仿真重点实验室开放课题基金资助(UPDMHITSZ2014B06)