面向复杂路网低频 GPS 采样数据新型地图匹配算法.

The number of low-frequency sampling traffic trajectory data is growing explosively in big data era. Map matching techniques play an essential role in taxi pick-up hot spot and route recommendation, which can help accurately and efficiently matching low-frequency floating car data from complex road networks. In view of the aforementioned discussion, this paper proposed an improved algorithm based on curve fitting. This algorithm filled missing trajectory data and road network data by interpolation and mean methods, employed Geohash technology to store and search road network data and trajectory data, took into full consideration the factors of the speed of vehicles and road speed limit, used the backward vector of trajectory point and road segment vector to analyze the candidate road segments, and designed a comprehensive evaluation function to obtain the optimal matching results. The experimental results show that the proposed curve-fitting algorithm has a higher matching accuracy and faster time efficiency when compared with the traditional vertical projection and curve -fitting algorithms. [ABSTRACT FROM AUTHOR]