An improved map-matching algorithmbased on lowfrequency GPS data

Abstract

It is important to match vehicle GPS points onto a map. Many map-matching algorithms, which use heading, speed and position to complete the map-matching process, are not suitable for the low frequency GPS data. The distance between two adjacent GPS points in the low frequency GPS data can range from 500m to 1000m, and the vehicle can cross several streets. This paper puts forward a modified weight-based algorithm using the vertical distance and the shortest path. This weight-based algorithm uses the instantaneous speed of two points to estimate the vehicle travel distance. Therefore, the improved algorithm has the higher accuracy.