Uni-PrevPredMap: Extending PrevPredMap to a Unified Framework of Prior-Informed Modeling for Online Vectorized HD Map Construction
Nan Peng, Xun Zhou, Mingming Wang, Guisong Chen, Wenqi Xu · 2025
Read Uni-PrevPredMap: Extending PrevPredMap to a Unified Framework of Prior-Informed Modeling for Online Vectorized HD Map Construction by Nan Peng, Xun Zhou, Mingming Wang, Guisong Chen, Wenqi Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-06647.txt.
Description
Nan Peng and co-authors fuse two kinds of prior maps for self-driving cars in a 2025 paper. Cars build high-definition road maps online, frame by frame, as they drive. Past predictions and stored map data both help, but each source can be wrong on its own.
Their framework, Uni-PrevPredMap, treats both sources as flawed priors in one shared vector form. A tri-mode training plan shows the model road scenes with no prior, with a time prior, and with both fused. One model thus stays safe with no help and gains when help is sound. Costly stored maps can then be used with care, not blind trust. Safer mapping is the end goal.
Catalog record
| CREATOR | Nan Peng, Xun Zhou, Mingming Wang, Guisong Chen, Wenqi Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-06647