Putting the car on the map: A scalable map matching system for the open source community

Abstract

Recent years have seen a proliferation of mobile devices connected to the internet, including connected cars. These systems send a stream of position annotated messages when requesting location based services. The position information in turn can be used to improve those services. Here, we focus on online map matching of the most recent position of a connected vehicle on a road map. This information, aggregated and privacy aware, can serve as a basis e.g. for machine learning algorithms used to improve traffic prediction. We describe a system for online map matching in the backend that implements a state of the art algorithm based on a Hidden Markov Model. This system uses only open source software and open data. The development of the map matcher was motivated by a perceived lack of a scalable system in the open source realm. We discuss its role as part of a scalable backend system designed to provide spatially aware services.