Constrained Diffusers for Safe Planning and Control

TL;DR

This paper introduces Constrained Diffusers, a framework that integrates safety constraints into diffusion-based planning models using constrained Langevin sampling, enabling safe and efficient control without retraining.

Contribution

It proposes a novel method to incorporate constraints into pre-trained diffusion models via iterative algorithms, ensuring safety without modifying the original model architecture.

Findings

Achieves constraint satisfaction with less computation time.

Demonstrates effectiveness in Maze2D, locomotion, and pybullet tasks.

Maintains competitive performance with existing methods.

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.