arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.01023cs.ROcs.AI
收藏

面向闭环自动驾驶的开源模块化扩散运动规划基准

An Open-Source Modular Benchmark for Diffusion-Based Motion Planning in Closed-Loop Autonomous Driving

Graduate School of Information Science and Technology, The University of Tokyo(信息科学与技术研究生学校,东京大学);TIER IV, Inc.(TIER IV公司)

Yun Li, Simon Thompson, Yidu Zhang, Ehsan Javanmardi, Manabu Tsukada

AI总结:

本研究提出开源模块化扩散运动规划基准,通过分解单体规划器并实现原生C++去噪循环,实现实车部署中运行时可配置的求解器参数和去噪过程的可观测性。

摘要:

基于扩散的运动规划器在nuPlan等基准上取得了最先进的结果,但其在闭环生产自动驾驶栈中的评估仍鲜有探索。现有评估抽象了ROS 2通信延迟和实时调度约束,而单体ONNX部署在导出时冻结所有求解器参数。我们提出一个开源模块化基准,解决了这两个缺口:使用ONNX GraphSurgeon,我们将单体18,398节点的扩散规划器分解为三个可独立执行的模块,并用原生C++重新实现DPM-Solver++去噪循环。集成到Autoware中的开源AD栈中,该系统在世界各地的实车部署中实现了运行时可配置的求解器参数,无需模型重新编译,并实现了每步去噪过程的可观察性,打破了单体部署的黑箱。与CARLA等独立模拟器的评估不同,我们的基准在生产级栈中运行,并通过AWSIM闭环模拟验证。通过系统比较DPM-Solver++(一阶和二阶)和DDIM在六个步数配置(N在{3,5,7,10,15,20})下的表现,我们显示编码缓存可减少3.2倍的延迟,二阶求解可将FDE在N=3时降低41%。完整的代码库将作为开源发布,为从模拟基准到实车部署提供直接路径。

英文摘要:

Diffusion-based motion planners have achieved state-of-the-art results on benchmarks such as nuPlan, yet their evaluation within closed-loop production autonomous driving stacks remains largely unexplored. Existing evaluations abstract away ROS 2 communication latency and real-time scheduling constraints, while monolithic ONNX deployment freezes all solver parameters at export time. We present an open-source modular benchmark that addresses both gaps: using ONNX GraphSurgeon, we decompose a monolithic 18,398 node diffusion planner into three independently executable modules and reimplement the DPM-Solver++ denoising loop in native C++. Integrated as a ROS 2 node within Autoware, the open-source AD stack deployed on real vehicles worldwide, the system enables runtime-configurable solver parameters without model recompilation and per-step observability of the denoising process, breaking the black box of monolithic deployment. Unlike evaluations in standalone simulators such as CARLA, our benchmark operates within a production-grade stack and is validated through AWSIM closed-loop simulation. Through systematic comparison of DPM-Solver++ (first- and second-order) and DDIM across six step-count configurations (N in {3, 5, 7, 10, 15, 20}), we show that encoder caching yields a 3.2x latency reduction, and that second-order solving reduces FDE by 41% at N=3 compared to first-order. The complete codebase will be released as open-source, providing a direct path from simulation benchmarks to real-vehicle deployment.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑