Studie představuje matematický popis minimální požadované přesnosti HD map pro autonomní vozidla, identifikuje klíčové faktory ovlivňující kvalitu map a nabízí nový vzorec určující nutnou přesnost jak pro statické, tak pro dynamické modely.
Abstract:
Autonomous vehicles rely on a combination of sensors for safe navigation around the world. For precise localisation, high-definition (HD) maps are used. These maps are a representation of the world containing information about objects on the road infrastructure. Currently, there are tens of HD map makers, however, no rigorous description of the requirements for the accuracy of HD maps has been published yet. This study fills the gap and offers a mathematical description of the minimum required accuracy for HD maps. In the first part, we identify factors that influence the quality of a map. Based on that, we proceed to present our solution for determining the minimum required accuracy for HD maps, both for static and dynamic models, and present a new formula for the minimum necessary accuracy for HD maps.
Autoři: KŘEHLÍK, Štěpán, Marek VANŽURA a Adam SKOKAN. Minimum required accuracy for HD maps. Journal of Navigation [online]. 1-17 [cit. 2023-07-24]. ISSN 0373-4633. Dostupné z: doi:10.1017/S0373463323000140.