Time-Unified Diffusion Policy with Action Discrimination for Robotic Manipulation

TL;DR

This paper introduces the Time-Unified Diffusion Policy (TUDP), a novel approach that unifies the denoising process in diffusion models for robotic manipulation, leading to faster and more accurate action generation.

Contribution

The paper proposes a time-unified diffusion policy with action discrimination, improving efficiency and accuracy in robotic manipulation by unifying the denoising process and incorporating action recognition.

Findings

Achieves state-of-the-art success rates on RLBench tasks.

Faster action generation with fewer denoising iterations.

Effective in real-world robotic tasks.

Abstract

In many complex scenarios, robotic manipulation relies on generative models to estimate the distribution of multiple successful actions. As the diffusion model has better training robustness than other generative models, it performs well in imitation learning through successful robot demonstrations. However, the diffusion-based policy methods typically require significant time to iteratively denoise robot actions, which hinders real-time responses in robotic manipulation. Moreover, existing diffusion policies model a time-varying action denoising process, whose temporal complexity increases the difficulty of model training and leads to suboptimal action accuracy. To generate robot actions efficiently and accurately, we present the Time-Unified Diffusion Policy (TUDP), which utilizes action recognition capabilities to build a time-unified denoising process. On the one hand, we build a time-unified velocity field in action space with additional action discrimination information. By unifying all timesteps of action denoising, our velocity field reduces the difficulty of policy learning and speeds up action generation. On the other hand, we propose an action-wise training method, which introduces an action discrimination branch to supply additional action discrimination information. Through action-wise training, the TUDP implicitly learns the ability to discern successful actions to better denoising accuracy. Our method achieves state-of-the-art performance on RLBench with the highest success rate of 82.6% on a multi-view setup and 83.8% on a single-view setup. In particular, when using fewer denoising iterations, TUDP achieves a more significant improvement in success rate. Additionally, TUDP can produce accurate actions for a wide range of real-world tasks.