arXiv：cs.LG· Ahmed Abouelazm, Rupert Polley, Qingyuan Zhang, Yin Wu, Philip Sch\"orner, Carl Esselborn, J. Marius Z\"ollner·· 6 小时前AI 评分38
超越路点回归：面向端到端驾驶的可达自车未来查询式代价学习
Beyond Waypoint Regression: Query-Based Cost Learning over Reachable Ego Futures for End-to-End Driving
AI 导读
研究者提出一种查询式代价学习框架，为动态可达的自车轨迹查询估计有界代价，而非依赖密集 BEV 网格或少量回归轨迹。该方法在 nuScenes 上优于 ST-P3、NMP 等代价估计规划器，在碰撞率上超过多数回归基线且 L2 保持竞争力；在真实驾驶日志中无需微调即比 SparseDrive 和 Alpamayo 降低碰撞率。该工作已被 ACCV 2026 接收。
正文
Abstract:End-to-end planners based on waypoint regression achieve strong open-loop accuracy, but they primarily learn to mimic expert geometry and remain difficult to adapt to deployment-time safety constraints. We propose a query-based cost-learning framework that estimates bounded costs for dynamically reachable ego trajectory queries, rather than dense BEV cells or a small regressed trajectory set. Compact joint scene tokens capture coherent multimodal agent futures, while contingency-aware cost aggregation and cost-guided intra-cluster MPPI mixing convert the learned cost topology into feasible ego plans. On nuScenes, our method improves over prior cost-estimation planners such as ST-P3 and NMP, outperforms most regression baselines in collision rate, while remaining competitive in L2, and retaining an interpretable cost interface. On real-world driving logs, the proposed planner reduces collision rates compared with SparseDrive and Alpamayo without fine-tuning, while maintaining a diverse set of candidate trajectories.
| Comments: | Accepted in the 18th Asian Conference on Computer Vision (ACCV 2026) |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Systems and Control (eess.SY) |
| Cite as: | arXiv:2610.08123 [cs.RO] |
| (or arXiv:2610.08123v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08123 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Ahmed Abouelazm [view email]
[v1]
Tue, 6 Oct 2026 10:43:13 UTC (27,436 KB)
来源：arXiv：cs.LG · arxiv.org