Personalised route-map generation using crowd sourced GPS traces

Abstract

Developing a robust and well-informed cellular phone based transportation system comes with increasing popularity and availability of digital maps that has certainly made navigation simpler. But that comes with a serious limitation of narrow unknown paths which does not hold any place on these maps. The drawbacks of inaccuracy of road geometry and the lack of fine-grained information hampers a route guidance system. In this paper, we present an approach to build high-precision maps from the GPS traces of crowd, equipped with differential GPS receivers, for updating existing paths and adding new path segments scaled according to digital map. Due to decreasing cost, we expect that large volume of GPS traces can be gathered for building up new road segments. Our approach is to build a smart system where the crowd traces will be uploaded on the server through a cellular user-interface, to generate a fine-grained map based on human labelling.