论文arXiv 2611.251382 天前
传统规划器在HUGSIM评测中超越UniAD端到端模型
本文由 AI 根据公开内容整理,可能有误,请以原文为准:Old-School Planners Drive (Sometimes) Better Than Learned Ones(arXiv 2611.25138)
核心内容
针对端到端自动驾驶模型是否完全取代传统规划器的争议,研究团队实现了16种经典规划算法组合,并在nuScenes、nuPlan及HUGSIM三大基准上进行严格评测。实验显示,在冻结感知输入且无高精地图条件下,最佳传统规划器在nuScenes开环测试中L2误差仅0.735m,优于UniAD的0.947m;在nuPlan Test14-hard闭环测试中得分79.3,超过规则基线PDM-Closed的76.1分,证明传统方案在标准学术基准上仍具强竞争力。
在更严苛的HUGSIM照片级真实感闭环仿真中,9/16的传统规划器得分超越UniAD学习式规划器。其中“Hybrid A*路径搜索+非线性轨迹优化”组合将HD-Score从32.0大幅提升至41.8,增益主要来自对静止及迎面车辆的避让;但在应对切入车辆时优势消失。归因分析表明,UniAD发生的182次碰撞中有31次可通过替换为最佳传统规划器避免,确认这些失效源于规划模块而非感知缺陷,为具身智能系统的模块化安全兜底提供了实证依据。
关键看点
- 性能反超:最佳传统规划器在HUGSIM闭环评测中HD-Score达41.8,显著高于UniAD端到端模型的32.0分
- 最优组合:Hybrid A*路径搜索结合非线性轨迹优化表现最佳,在nuScenes开环L2误差降至0.735m
- 失效归因:UniAD的182次碰撞事故中31次系规划器缺陷导致,替换传统算法即可成功完成路线
- 场景局限:传统规划器对静止和迎面车辆处理更优,但对动态切入车辆的博弈能力未见提升
工程与产业观察
该研究揭示了当前端到端模型在长尾安全场景下的规划短板,验证了“感知端到端+规划模块化”混合架构的工程价值。对于具身智能开发者,在部署纯神经网络策略前,保留基于Hybrid A*与非线性优化的传统规划器作为安全冗余或蒸馏教师模型,仍是提升系统鲁棒性与可解释性的有效路径。
英文摘要 (Abstract)
End-to-end driving models learn to plan from data, and hand-written planners are widely treated as a thing of the past. We test whether this is justified: we implement sixteen textbook planners, each combining a path source (reference line, A*, hybrid A*, RRT*) with a trajectory method (Frenet lattice, nonlinear optimization, $s$--$t$ search, dynamic window), and evaluate them on three benchmarks: open loop on nuScenes, closed loop on nuPlan with ground-truth perception, and photorealistic closed loop on HUGSIM. On HUGSIM, they replace only the planner of the end-to-end model UniAD and read its frozen perception, without an HD map or other privileged information. On nuScenes and nuPlan, the best textbook planners are competitive (L2 error of 0.735,m against 0.947,m for UniAD's planner; a nuPlan score of 79.3 against 76.1 for the rule-based PDM-Closed on Test14-hard with reactive agents). On HUGSIM, nine of the sixteen score above UniAD's learned planner on the same perception,. The best, a hybrid A* path followed by nonlinear trajectory optimization, raises HUGSIM's driving score (HD-Score) from 32.0 to 41.8. The gain is largest with standing and oncoming cars; against cars that steer at the ego, there is none. In 31 of the 182 runs in which UniAD's planner hits a car, the best textbook planner completes the route on the same perception: these crashes are due to the planner, not the perception.
作者:Andreas Look
外部来源
本页事实对应的公开页面,以原文为准。(点击将离开本站)