论文部分内容阅读
针对基于浮动车辆数据(floating car data,FCD)的城市道路交通信息采集系统存在的问题,提出一种基于最小二乘支持向量机(LS-SVM)和证据理论的数据融合方法,通过融合地感线圈采集的交通流量信息,提高FCD系统交通速度信息采集的准确性.利用LS-SVM回归得到速度-流量关系曲线的临界速度参数,再根据历史数据库用统计方法计算出流量-速度关联规则的可信度矩阵,在得到这些经验知识的基础上,定义了两种证据源的基本概率分配函数.最后,通过D-S证据理论对两种证据源进行数据融合,获得融合后的速度信息.实地跑车实验结果论证了融合算法的有效性和可靠性.
Aiming at the existing problems of urban road traffic information collection system based on floating car data (FCD), a data fusion method based on LS-SVM and evidence theory is proposed. By means of fusion The traffic flow information collected by the coil is used to improve the accuracy of the traffic speed information acquisition of the FCD system.The critical speed parameters of the speed-flow relation curve are obtained by LS-SVM regression and the traffic-speed association rules are calculated by the statistical method according to the historical database Based on these empirical knowledge, the basic probability distribution function of two kinds of evidence sources is defined.Finally, the data of two kinds of evidence sources are fused by DS evidence theory to obtain the fused speed information. The results demonstrate the effectiveness and reliability of the fusion algorithm.