Research

FA-RDP: Diffusion Policies That Switch Frequency When Contact Hits

Shar Hendrix 4 min read

Contact-rich robot tasks have a split personality. Before touch, several approach trajectories can be equally valid. After contact, force feedback has to be fast or the grasp fails. FA-RDP (Frequency-Adaptive Reactive Diffusion Policy) is a new method from Shanghai Jiao Tong University, Shanghai Innovation Institute, and Noematrix that lets one diffusion policy switch sampling rate mid-episode instead of picking one frequency forever.

FA-RDP teaser showing multimodal pre-contact paths and high-frequency contact responses
FA-RDP teaser: multimodal approaches before contact, reactive force control after. Source: FA-RDP project page.

The tradeoff standard diffusion policies lose

The authors’ project page and arXiv abstract (arXiv:2607.28596) put the problem cleanly. Fixed low-frequency, multi-step diffusion sampling preserves diverse pre-contact modes but reacts slowly to force. High-frequency sampling improves reactivity but collapses distinct approach modes.

FA-RDP’s answer:

  • A shared multi-frequency visual-force Transformer that can predict action chunks at low or high frequency
  • A learned multimodality indicator that selects multi-step low-frequency sampling before contact and one-step high-frequency sampling as ambiguity drops
  • Manifold Consistency Distillation (MCD) so the high-frequency path can run as a one-step sampler on the robot’s action manifold

In their framing, low frequency runs around 10 Hz for approach; high frequency around 30 Hz once contact demands force response.

FA-RDP multi-frequency visual-force Transformer architecture diagram
Multi-frequency Transformer with dense (30 Hz) and sparse (10 Hz) temporal grids. Source: FA-RDP project page.

Numbers on three contact-rich tasks

On box flipping, switch toggling, and button pressing (20 trials each), the project reports:

MethodAvg. success
Diffusion Policy (DP)10.0%
RDP35.0%
ImplicitRDP51.7%
Regression + force20.0%
FA-RDP81.7%

Indicator-guided switching lifts average success from 61.7% (always high-frequency) to 81.7%. The authors also show the low-frequency path covers multiple approach modes that a high-frequency-only policy collapses to one.

Code is listed as coming soon on the project site.

A Human’s Take

This is the kind of paper I want more of: not “diffusion, but bigger,” but “diffusion, but the control loop matches the physics phase.” Contact is where factories lose money and demos die.

I’ll believe the 81.7% harder once the code lands and other labs rerun the three tasks. Until then, the core idea—slow and multimodal before touch, fast and distilled after—is already worth stealing.

Sources