RA-DP: Rapid Adaptive Diffusion Policy for Training-Free High-frequency Robotics Replanning

2025-03-06

AI Summary

Diffusion models exhibit strong generalization in robotic tasks but struggle in highly dynamic environments due to their iterative denoising mechanism, which inherently limits replanning frequency (hindering real-time responsiveness) or compromises adaptability (failing to incorporate unexpected sensory feedback). To address this, we propose a training-agnostic adaptive diffusion policy framework. Its core innovations are: (i) an action queue buffer and scheduling mechanism, and (ii) online integration of environment-guided signals into each denoising step, enabling millisecond-scale action updates at up to 100 Hz. The method requires no fine-tuning or retraining and is compatible with arbitrary external guidance signals. Evaluated in both simulation and real-robot experiments, it significantly improves success rates on dynamic tasks and replanning latency, consistently outperforming state-of-the-art diffusion-based policies.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Diffusion models exhibit impressive scalability in robotic task learning, yet they struggle to adapt to novel, highly dynamic environments. This limitation primarily stems from their constrained replanning ability: they either operate at a low frequency due to a time-consuming iterative sampling process, or are unable to adapt to unforeseen feedback in case of rapid replanning. To address these challenges, we propose RA-DP, a novel diffusion policy framework with training-free high-frequency replanning ability that solves the above limitations in adapting to unforeseen dynamic environments. Specifically, our method integrates guidance signals which are often easily obtained in the new environment during the diffusion sampling process, and utilizes a novel action queue mechanism to generate replanned actions at every denoising step without retraining, thus forming a complete training-free framework for robot motion adaptation in unseen environments. Extensive evaluations have been conducted in both well-recognized simulation benchmarks and real robot tasks. Results show that RA-DP outperforms the state-of-the-art diffusion-based methods in terms of replanning frequency and success rate. Moreover, we show that our framework is theoretically compatible with any training-free guidance signal.

Problem

• Adapts to novel, highly dynamic robotic environments
• Enables high-frequency replanning without retraining
• Integrates guidance signals for improved motion adaptation

Innovation

• Training-free high-frequency replanning framework
• Guidance signals integrated during diffusion sampling
• Action queue mechanism for rapid replanning