基于曲线拟合和网络拓扑的综合地图匹配算法

A Comprehensive Map Matching Algorithm Based on Curve Fitting and Network Topology

为了使车辆导航系统能给人们提供精确可靠的车辆位置信息,提出了1种减少车辆定位误差的综合地图匹配算法。该算法将基于曲线拟合和网络拓扑关系的2种地图匹配算法结合起来,采用误差圆法、最小二乘法及投影法,解决基于曲线拟合算法在靠近的平行路段匹配率不佳的问题。用北京市物流电动车的GPS数据进行多个样本的实验验证,匹配率均在95%以上,单点平均匹配时间在5ms以内,与基于曲线拟合的地图匹配算法相比,综合算法复杂度降低,准确度和实时性提高。结果表明,综合算法可以应用到实际地图匹配工作中,尤其对于相互靠近的平行路段有明显的优势。

In order to provide accurate and reliable information of vehicle location in vehicle navigation system, a comprehensive map matching algorithm to reduce error is presented. The algorithm combines the curve fitting with the network topology information to reduce the error of vehicle location. Several analysis methods, including Error Circle, Least Square and Projection, are adopted to solve the low match rate on the parallel roads based on curve fitting method. The GPS data from Beijing logistics electric vehicles are used to verify the algorithm. The results show that the match rate is above 95% and the match time is within 5ms. Compared with the curve fitting algorithm, the comprehensive algorithm decreases the calculation complexity, improves the accuracy and real-time performance, and can be applied to the actual map matching work. The algorithm especially has obvious advantages for parallel sections close to each other.