TacDiffusion: Force-domain Diffusion Policy for Precise Tactile Manipulation

TL;DR

This paper introduces TacDiffusion, a diffusion model-based framework for high-precision tactile robotic insertion, achieving high transfer success and robustness through a novel dynamic filtering approach.

Contribution

It presents a new diffusion model framework for tactile manipulation that transfers skills zero-shot and improves success with dynamic filtering, addressing frequency misalignment issues.

Findings

Achieves 95.7% zero-shot transfer success rate across various tasks.

Improves task success rate by 9.15% using dynamic system-based filtering.

Provides guidelines on balancing inference speed and model capability.

Abstract

Assembly is a crucial skill for robots in both modern manufacturing and service robotics. However, mastering transferable insertion skills that can handle a variety of high-precision assembly tasks remains a significant challenge. This paper presents a novel framework that utilizes diffusion models to generate 6D wrench for high-precision tactile robotic insertion tasks. It learns from demonstrations performed on a single task and achieves a zero-shot transfer success rate of 95.7% across various novel high-precision tasks. Our method effectively inherits the self-adaptability demonstrated by our previous work. In this framework, we address the frequency misalignment between the diffusion policy and the real-time control loop with a dynamic system-based filter, significantly improving the task success rate by 9.15%. Furthermore, we provide a practical guideline regarding the trade-off between diffusion models' inference ability and speed.