Constrained Diffusers for Safe Planning and Control
Abstract¶
Diffusion models have shown remarkable potential in planning and control tasks due to their ability to represent multimodal distributions over actions and trajectories. However, ensuring safety under constraints remains a critical challenge for diffusion models. This paper proposes Constrained Diffusers, a novel framework that incorporates constraints into pre-trained diffusion models without retraining or architectural modifications. Inspired by constrained optimization, we apply a constrained Langevin sampling mechanism for the reverse diffusion process that jointly optimizes the trajectory and realizes constraint satisfaction through three iterative algorithms: projected method, primal-dual method and augmented Lagrangian approaches. In addition, we incorporate discrete control barrier functions as constraints for constrained diffusers to guarantee safety in online implementation. Experiments in Maze2D, locomotion, and pybullet ball running tasks demonstrate that our proposed methods achieve constraint satisfaction with less computation time, and are competitive to existing methods in environments with static and time-varying constraints.
Summary¶
Zhang, Zhao, Papachristodoulou, and Umenberger adapt pretrained diffusion planners to constrained planning and control by modifying the reverse sampling process rather than retraining the model. Constrained Diffusers use projected, primal-dual, and augmented-Lagrangian sampling updates, and the paper connects those updates with discrete control barrier functions for receding-horizon safety.