Tsukada Lab. Computer Network Group
← Publications Int'l conference

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

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

29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026), Naples, Italy, Sep 15, 2026

Summary in three lines

  • This open benchmark runs modular diffusion motion planning inside an Autoware ROS 2 closed loop.
  • It splits a monolithic model and moves the denoising solver to C++, enabling caching and runtime solver choices.
  • AWSIM tests quantify the latency gains of caching and the trajectory-error reduction of a second-order solver.

Figures from the paper

Figure 1: Monolithic and modular diffusion planners
Figure 1. Monolithic and modular diffusion plannersSource: Figure 1 from Li et al., 29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026) (2026) (authors’ figure from the manuscript)The modular implementation runs the encoder once and reuses its context during denoising steps.
Figure 4: Latency–accuracy trade-off
Figure 4. Latency–accuracy trade-offSource: Figure 4 from Li et al., 29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026) (2026) (authors’ figure from the manuscript)The frontier compares encoder caching and solver choices against a 100 ms planning budget.

Abstract

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. Project page:

Related news

How to cite

Y. Li, S. Thompson, Y. Zhang, E. Javanmardi, and M. Tsukada, “An Open-Source Modular Benchmark for Diffusion-Based Motion Planning in Closed-Loop Autonomous Driving,” in 29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026), Naples, Italy, 2026.

BibTeX

@inproceedings{Yun2026,
  title         = {An Open-Source Modular Benchmark for Diffusion-Based Motion Planning in Closed-Loop Autonomous Driving},
  author        = {Yun Li and Simon Thompson and Yidu Zhang and Ehsan Javanmardi and Manabu Tsukada},
  url           = {https://arxiv.org/abs/2603.01023},
  year          = {2026},
  date          = {2026-09-15},
  urldate       = {2026-09-15},
  booktitle     = {29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026)},
  address       = {Naples, Italy},
  abstract      = {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. Project page:},
  keywords      = {ITS},
  pubstate      = {published},
  tppubtype     = {inproceedings}
}