LGmap: Local-to-Global Mapping Network for Online Long-Range Vectorized HD Map Construction
This work addresses the challenge of mapless driving for autonomous vehicles, representing an incremental improvement with specific gains in stability and convergence.
The authors tackled the problem of online long-range vectorized HD map construction for autonomous driving by introducing LGmap, a pipeline that achieved a UniScore of 0.66 on the Mapless Driving OpenLaneV2 test set.
This report introduces the first-place winning solution for the Autonomous Grand Challenge 2024 - Mapless Driving. In this report, we introduce a novel online mapping pipeline LGmap, which adept at long-range temporal model. Firstly, we propose symmetric view transformation(SVT), a hybrid view transformation module. Our approach overcomes the limitations of forward sparse feature representation and utilizing depth perception and SD prior information. Secondly, we propose hierarchical temporal fusion(HTF) module. It employs temporal information from local to global, which empowers the construction of long-range HD map with high stability. Lastly, we propose a novel ped-crossing resampling. The simplified ped crossing representation accelerates the instance attention based decoder convergence performance. Our method achieves 0.66 UniScore in the Mapless Driving OpenLaneV2 test set.