DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving

TL;DR

DiffusionDrive introduces a truncated diffusion model with multi-mode anchors and an efficient decoder, enabling real-time, diverse, and high-quality autonomous driving actions with fewer denoising steps.

Contribution

It proposes a novel truncated diffusion policy with multi-mode anchors and an efficient cascade decoder for end-to-end autonomous driving.

Findings

Achieves 10× reduction in denoising steps compared to vanilla diffusion policies.

Sets a new record of 88.1 PDMS on NAVSIM dataset.

Runs at 45 FPS on NVIDIA 4090, demonstrating real-time performance.

Abstract

Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10$\times$ reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions. Code and model will be available at https://github.com/hustvl/DiffusionDrive.