一种基于相关性分析的地图匹配方法

A map matching algorithm based on correlation analysis

为了改善机动车导航地图匹配的精度,提出了一种新的基于递推轨迹与候选路径相关性分析的地图匹配方法.为了获得最佳匹配路径,分别计算递推轨迹曲线和候选路径曲线的切向量角,并沿曲线分段计算切向量角的一阶差分,剔除机动车递推轨迹曲线与候选路径曲线间的相对转角,将二维曲线相关问题降为一维曲线的相关问题.基于差分切向量角序列,计算递推轨迹曲线与不同候选路径曲线的相关系数,获得最佳匹配路径,定义具有最大相关系数的路径段为最佳匹配段.实例说明,新的地图匹配的方法对相似候选路径具有正确的分辨能力,对递推数据的随机误差有较强的鲁棒性.

A map matching method based on correlation analysis of dead reckoning tracks and candidate routes was proposed in order to improve map matching precision in vehicle navigation. To find the best matching route, the curve tangent vector angles of the tracks and candidate routes were calculated and one-order differentiated along the different sections of the curves. By eliminating the relative rotation angle of tracks and all candidate route curves, the 2D-curve related problems were simplified to one dimension. The best match route were determined based on the correlation coefficients from the tangent vector angle series after first order difference, and the route segment corresponding to the maximum correlation coefficient was defined as the best match segment （BMS）. A case study shows that the map matching method can differentiate similar candidate routes correctly and has high robustness in terms of random error of the dead reckoning data.