FlexMap Fusion: Georeferencing and Automated Conflation of HD Maps with OpenStreetMap

TL;DR

FlexMap Fusion tackles the challenge of integrating OpenStreetMap data into existing HD vector maps to reduce manual HD map generation effort. The authors introduce a three-module pipeline—Map Alignment, Map Conflation, and Georeferencing—that first aligns a LiDAR-based PCM/VM to an RTK-corrected GNSS trajectory, then conflates semantic attributes from OSM into the Lanelet2-based VM, and finally georeferences the result to a global frame. Key contributions include a practical, open-source workflow compatible with Autoware, a robust matching approach using a Buffer Growing algorithm, and validated georeferencing with real-world EDGAR data, achieving sub-meter alignment under favorable conditions. The work demonstrates increased map completeness and scalability for autonomous driving stacks, while highlighting limitations in complex road structures and the need for automatic control-point selection and up-to-date map data for broader deployment.

Abstract

Today's software stacks for autonomous vehicles rely on HD maps to enable sufficient localization, accurate path planning, and reliable motion prediction. Recent developments have resulted in pipelines for the automated generation of HD maps to reduce manual efforts for creating and updating these HD maps. We present FlexMap Fusion, a methodology to automatically update and enhance existing HD vector maps using OpenStreetMap. Our approach is designed to enable the use of HD maps created from LiDAR and camera data within Autoware. The pipeline provides different functionalities: It provides the possibility to georeference both the point cloud map and the vector map using an RTK-corrected GNSS signal. Moreover, missing semantic attributes can be conflated from OpenStreetMap into the vector map. Differences between the HD map and OpenStreetMap are visualized for manual refinement by the user. In general, our findings indicate that our approach leads to reduced human labor during HD map generation, increases the scalability of the mapping pipeline, and improves the completeness and usability of the maps. The methodological choices may have resulted in limitations that arise especially at complex street structures, e.g., traffic islands. Therefore, more research is necessary to create efficient preprocessing algorithms and advancements in the dynamic adjustment of matching parameters. In order to build upon our work, our source code is available at https://github.com/TUMFTM/FlexMap_Fusion.

Figures (9)

• Figure 1: PCM with original VM (bottom image) and exemplary information from OSM to be conflated (top image).
• Figure 2: Overview of the presented architecture of FlexMap Fusion.
• Figure 3: Preprocessing of Lanelet2 map.
• Figure 4: Principle of Buffer Growing algorithm.
• Figure 5: Attribute transfer correspondences between OSM (orange) and Lanelet2 (blue).