Problem
Research questions and friction points this paper is trying to address.
Imitation Learning
Closed-Loop Robot Control
Prediction Latency
Diffusion Policy
Flow Matching
Adaptive Mean Flow for Responsive Closed-Loop Robot Control
🤖 AI Summary
This study addresses the high latency and closed-loop control jitter caused by iterative denoising in diffusion and flow models for robotic manipulation. We propose an adaptive Mean Flow-based robot control method that integrates imitation learning with Mean Flow to accelerate Flow Matching, thereby enhancing real-time performance. Furthermore, a progressive signal-to-noise ratio mechanism is introduced to predict actions based on preceding trajectory steps, effectively suppressing control discontinuities while preserving adaptability to future dynamics. Experimental evaluations demonstrate that the proposed approach significantly outperforms existing baselines in both simulated and real-world tasks. By effectively balancing real-time responsiveness with control stability, our method substantially improves closed-loop control performance, offering a robust solution for deploying generative policies in dynamic robotic environments.
📝 Abstract
Diffusion- and flow-based robot policies have recently become widespread in robotic Imitation Learning (IL) due to their high performance and ability to model continuous and multimodal distributions. However, the iterative denoising procedure used by these models introduces significant prediction latency, hindering high-frequency closed-loop robot control and leading to jittery, unstable motion when frequent updates to the robot’s action predictions are used. Therefore, it is common practice to train models to predict chunks of actions that can be executed sequentially without feedback, even when this reduces responsiveness and may mean the most recent state information is not used. In this article, we present Adaptive Mean Flow (AMF), a flow-based IL method that enables smooth and responsive, fully closed-loop robot control. AMF uses Mean Flow, which is an accelerated form of Flow Matching (FM), to minimize prediction latency. To ensure smoothness and consistency across predictions, AMF uses a corrupted version of the trajectory from the previous step when predicting new robot actions, with the signal-to-noise ratio increasing over the time parameter of the trajectory. This discourages large changes in the prediction from one step to the next, while allowing freedom to adapt the predictions for future steps. We evaluate AMF across a wide range of simulated and real robot tasks and demonstrate significantly improved performance compared with baselines.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Adaptive Mean Flow
Flow Matching
Closed-Loop Robot Control
Imitation Learning
Prediction Latency
🔎 Similar Papers
No similar papers found.
