实时分布式地图匹配系统的设计与实现

Design and Realization of a Real-Time Distributed Map-Matching System

在智能交通系统的基础上,设计实现了基于实时浮动车数据的分布式地图匹配系统。分布式架构的应用旨在提高系统数据接收和处理的实时性。同时提出了一种基于道路网格和最短路径的快速地图匹配算法,通过网格划分确定候选路段,根据二次网格划分、GPS角度和数量对候选路段权重进行修正,并利用最短路径算法寻求最符合路径,在保证处理效率的同时匹配的精度也有所提升。仿真实验选用三台匹配处理节点,经验证,完成大连市1000辆出租车约130万个GPS点的地图匹配用时为1分37秒,算法的平均正确匹配率为93.3%。实验结果满足实际智能交通系统的基本要求。

Based on the intelligent transportation system, this paper designs and realizes a distributed map-matching system for real-time probe vehicle data. The distributed structure aims at improving the real-time performance of data receiving and processing. This paper also proposes a quick map-matching al- gorithm on the basis of road meshing and the shortest path routing. The algorithm uses road meshing method to decide the candidate roads, revises the weight parameter according to the second road meshing, GPS angle and amount, then searches the best path according to the shortest path routing algorithm. The mapmatching precision is improved while the processing efficiency is guaranteed. This paper uses three distributed map-matching nodes to simulate. The result shows that it takes 97 seconds to finish the processing of 1300 thousand GPS data from 1000 probe vehicles in Dalian and the correct map-matching rate is 93.3% on average. The experimental results satisfy the basic requirement of intelligent transportation system.