AMap: Distilling Future Priors for Ahead-Aware Online HD Map Construction

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Li, Ruikai, Li, Xinrun, Xie, Mengwei, Shan, Hao, Qiu, Shoumeng, Chang, Xinyuan, Fan, Yizhe, Xiong, Feng, Jiang, Han, Ren, Yilong, Yu, Haiyang, Xu, Mu, Long, Yang, Ojha, Varun, Cui, Zhiyong
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918258978848768
author Li, Ruikai
Li, Xinrun
Xie, Mengwei
Shan, Hao
Qiu, Shoumeng
Chang, Xinyuan
Fan, Yizhe
Xiong, Feng
Jiang, Han
Ren, Yilong
Yu, Haiyang
Xu, Mu
Long, Yang
Ojha, Varun
Cui, Zhiyong
author_facet Li, Ruikai
Li, Xinrun
Xie, Mengwei
Shan, Hao
Qiu, Shoumeng
Chang, Xinyuan
Fan, Yizhe
Xiong, Feng
Jiang, Han
Ren, Yilong
Yu, Haiyang
Xu, Mu
Long, Yang
Ojha, Varun
Cui, Zhiyong
contents Online High-Definition (HD) map construction is pivotal for autonomous driving. While recent approaches leverage historical temporal fusion to improve performance, we identify a critical safety flaw in this paradigm: it is inherently ``spatially backward-looking." These methods predominantly enhance map reconstruction in traversed areas, offering minimal improvement for the unseen road ahead. Crucially, our analysis of downstream planning tasks reveals a severe asymmetry: while rearward perception errors are often tolerable, inaccuracies in the forward region directly precipitate hazardous driving maneuvers. To bridge this safety gap, we propose AMap, a novel framework for Ahead-aware online HD Mapping. We pioneer a ``distill-from-future" paradigm, where a teacher model with privileged access to future temporal contexts guides a lightweight student model restricted to the current frame. This process implicitly compresses prospective knowledge into the student model, endowing it with ``look-ahead" capabilities at zero inference-time cost. Technically, we introduce a Multi-Level BEV Distillation strategy with spatial masking and an Asymmetric Query Adaptation module to effectively transfer future-aware representations to the student's static queries. Extensive experiments on the nuScenes and Argoverse 2 benchmark demonstrate that AMap significantly enhances current-frame perception. Most notably, it outperforms state-of-the-art temporal models in critical forward regions while maintaining the efficiency of single current frame inference.
format Preprint
id arxiv_https___arxiv_org_abs_2512_19150
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AMap: Distilling Future Priors for Ahead-Aware Online HD Map Construction
Li, Ruikai
Li, Xinrun
Xie, Mengwei
Shan, Hao
Qiu, Shoumeng
Chang, Xinyuan
Fan, Yizhe
Xiong, Feng
Jiang, Han
Ren, Yilong
Yu, Haiyang
Xu, Mu
Long, Yang
Ojha, Varun
Cui, Zhiyong
Computer Vision and Pattern Recognition
Online High-Definition (HD) map construction is pivotal for autonomous driving. While recent approaches leverage historical temporal fusion to improve performance, we identify a critical safety flaw in this paradigm: it is inherently ``spatially backward-looking." These methods predominantly enhance map reconstruction in traversed areas, offering minimal improvement for the unseen road ahead. Crucially, our analysis of downstream planning tasks reveals a severe asymmetry: while rearward perception errors are often tolerable, inaccuracies in the forward region directly precipitate hazardous driving maneuvers. To bridge this safety gap, we propose AMap, a novel framework for Ahead-aware online HD Mapping. We pioneer a ``distill-from-future" paradigm, where a teacher model with privileged access to future temporal contexts guides a lightweight student model restricted to the current frame. This process implicitly compresses prospective knowledge into the student model, endowing it with ``look-ahead" capabilities at zero inference-time cost. Technically, we introduce a Multi-Level BEV Distillation strategy with spatial masking and an Asymmetric Query Adaptation module to effectively transfer future-aware representations to the student's static queries. Extensive experiments on the nuScenes and Argoverse 2 benchmark demonstrate that AMap significantly enhances current-frame perception. Most notably, it outperforms state-of-the-art temporal models in critical forward regions while maintaining the efficiency of single current frame inference.
title AMap: Distilling Future Priors for Ahead-Aware Online HD Map Construction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.19150