基于经验修正策略的延时地图匹配算法

An empirical correction strategy based delays map-matching algorithm

基于浮动车的交通流分析要求地图匹配快速而准确地处理GPS数据,现有的地图匹配算法无法满足交通流分析的实时处理要求。本算法采用延时策略,利用历史数据和最新数据及低速识别提取的静止点来指导延时点匹配,同时利用延时点来修正最新点的匹配经验。使用广州市路网及广东省交通厅浮动车的数据进行实验,结果证明在不同样本密度下,算法匹配正确率都可以达到80%以上,在精度、实时性和容错能力上均高于隐马尔科夫链匹配算法、时空匹配算法。

A map-matching algorithm is an algorithm that reflects the location based on Global Positioning Data（GPS） to the digital road network. Also,the floating vehicle based traffic flow analysis requires a map-matching algorithm to processing GPS data rapidly and accurately,but the existing map matching algorithms cannot meet the real-time GPS data processing requirement of traffic flow analysis.The empirical correction strategy based delays map-matching algorithm uses the delay strategy to match GPS data,which means that the algorithm employs historical data and latest data and the static point extracted by low-speed identification to guide the delay GPS data to the match,while taking advantage of the latest delay point to fix the match empirically of the latest point.Experiments were performed on the road network of Guangzhou and the flow cars data of the Guangdong Provincial Transportation Department.The results showed that the correct match rate of the algorithm could reach more than 80% at different sample densities and the accuracy and real time and fault tolerance of the algorithm were higher than the hidden Markov matching algorithm and spatial and temporal matching algorithm.