Framework of updating and interaction for high-definition maps with vehicle-road-cloud context

With the proposal of a Chinese solution for autonomous driving, high-definition (HD) maps have become an important infrastructure for achieving high-level autonomous driving. HD maps are vehicle-centric and are updated at an exceptionally rapid rate. In response to the weak correlation between static and dynamic data in HD maps and different update frequencies as well as the differences in data update modes between vehicle- and cloud-end HD maps, we propose a comprehensive, multilevel, and strongly associated HD map updating and interaction framework. The framework systematically proposes a logical relationship for dynamic HD map updating by integrating HD map models, collaborative update strategies, and the association of dynamic and static data. At the data-flow level, it achieves a rapid update and real-time interaction process of dynamic HD maps in the vehicle-road-cloud context. At the cloud end, we adopted a divide-and-conquer strategy to disseminate static data and real-time data. Meanwhile, HD maps were disseminated at the vehicle end, centered on the vehicle, through vehicle attributes and route-positioning constraints. The results showed that this framework can improve the driving efficiency and safety of autonomous vehicles in different scenarios, which can help develop digital roads and intelligent transportation on the Digital Earth.