SafeDiffuser: Safe Planning with Diffusion Probabilistic Models
About
Diffusion model-based approaches have shown promise in data-driven planning, but there are no safety guarantees, thus making it hard to be applied for safety-critical applications. To address these challenges, we propose a new method, called SafeDiffuser, to ensure diffusion probabilistic models satisfy specifications by using a class of control barrier functions. The key idea of our approach is to embed the proposed finite-time diffusion invariance into the denoising diffusion procedure, which enables trustworthy diffusion data generation. Moreover, we demonstrate that our finite-time diffusion invariance method through generative models not only maintains generalization performance but also creates robustness in safe data generation. We test our method on a series of safe planning tasks, including maze path generation, legged robot locomotion, and 3D space manipulation, with results showing the advantages of robustness and guarantees over vanilla diffusion models.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Robotic Path Planning | Maze2D (test) | BS1-0.003 | 22 | |
| PDE control | PDE control (test) | Rsample100 | 10 | |
| Robotic Manipulation | Robotic Manipulation Task (test) | Safety Rate10 | 10 | |
| Maze Navigation | Maze navigation task | Safety Rate0.00e+0 | 10 | |
| Closed-loop motion planning | nuPlan all-collision challenge set 68-scenario subset 14 (val) | Collision Rate78.57 | 7 |