测试时优化轨迹，让自动驾驶规划更准
Test-Time Trajectory Optimization for Autonomous Driving
- 用交叉熵法在测试时搜索最优轨迹，提升规划质量
- 在三个基准上平均提升56.3%以上，无需重新训练
- 可直接接入现有规划器，适合想改进决策的开发者
端到端自动驾驶规划器通常生成一组候选轨迹，评分后选择得分最高的。但评分器仅在轨迹生成后应用，无法影响轨迹集合本身，导致候选集质量受限时性能难以提升。本文提出TOAD方法，将评分器视为可学习的轨迹级奖励函数，在测试时通过交叉熵法搜索最大化该奖励的轨迹。该方法以原始规划器输出为初始解进行热启动，无需重训练，可即插即用。在六种基线规划器上，TOAD在NAVSIM-v1（94.7 PDMS）、NAVSIM-v2（56.3 EPDMS）和闭环HUGSIM基准上均取得显著提升。代码将通过项目页公开：https://valeoai.github.io/TOAD/。
原文摘要 · Abstract (English)
End-to-end planners for autonomous driving typically generate a set of candidate trajectories, score each one, and return the highest-scoring candidate. However, the scorer is applied only after the proposals are generated and cannot influence the set of trajectories: a weak set of candidates limits planning performance regardless of the scorer's quality. We instead treat the scorer as a learned trajectory-level reward function and search for trajectories that maximize it. Our method, TOAD, runs the Cross-Entropy Method at test time, warm-started from the planner's proposals. It requires no retraining and is plug-and-play for existing planners. Across six base planners, TOAD improves results on NAVSIM-v1 (94.7 PDMS), NAVSIM-v2 (56.3 EPDMS), and the closed-loop HUGSIM benchmark. The code will be made publicly available via the project page: https://valeoai.github.io/TOAD/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。