Introducing CrowdMapping: A Novel System for Generating Autonomous Driving Aiding Traffic Network Databases

Abstract

High definition maps of the road networks and the roads' environment have proven to be utterly useful for autonomous driving. Such maps can prove to be useful for the autonomous vehicle for numerous purposes - e.g. preliminary route planning, danger preparation and avoidance, etc. However, producing sufficient data for such maps can be costly because of the high variability of road conditions in the time domain and depending on the load of the elements of the given piece of transport infrastructure - i.e. road loads. In this paper, the CrowdMapping architecture is introduced, which presents a novel framework developed for the generation of an extensive and high definition road database, exploiting the opportunities offered by crowdsourcing, image processing, and cloud computing. State-of-the-art research is presented in the fields related to the development of the functions of the CrowdMapping framework. The currently ongoing research and development activities linked to CrowdMapping carried out at the Budapest University of Technology and Economics are also listed in chapter III. of this paper, as well as the future work possibilities, which are listed in chapter IV.