Validation of Mapping and Localization for Autonomous Vehicles (VAMLAV)

Important results from the project

The VAMLAV project primarily aimed to create and deliver a dataset that encompassed an HD-map of the Rural Road at AstaZero, along with sensor data collected under various weather conditions and seasons. Another objective was to develop methods for HD-map validation against reality, which involves linking the dataset with a known measurement uncertainty to a common reference frame. During the course of the project, requirements and specifications were established for suitable objects to be used as anchor points, and a static object was created at AstaZero for future testing purposes

Expected long term effects

"Within the framework of the VAMLAV project, we have created an extensive dataset that we expect will be utilized in future publications and research initiatives. We also hope that our research will significantly contribute to the development of crowdsourced HD-maps, especially in terms of validation methods and map creation processes. Through this work, we aim to lay the foundation for a more accurate and reliable use of HD-maps in the automotive industry

Approach and implementation

From the start, the VAMLAV project aimed to create a rich dataset from Rural Road at AstaZero. The pandemic and GDPR issues added further challenges to an already technically demanding area, but through collaboration between AstaZero, Zenseact, RISE, and AI Sweden, we overcame them. Focusing on HD-map validation, we linked the dataset to a common reference frame. AstaZero´s static objects enriched our project´s versatility and future prospects.