UniReflex: Plug-and-Play Force Control for Pretrained Generative Policies via Fast-Slow Reflex

2026-08-18

AI Summary

为了解决预训练生成策略缺乏力控制的问题，提出UniReflex框架，通过快速反射网络和自适应门控机制实现无需重新训练的力控制。

Abstract

Generative imitation learning policies excel at trajectory planning but lack closed-loop force regulation, while directly incorporating force modalities often requires redesigning or retraining the network. We present UniReflex, a universal plug-and-play framework that equips frozen generative policies with variable impedance control (VIC) for contact regulation, guided by force-direction intent collected during demonstration, without further slow-backbone fine-tuning. By non-invasively intercepting deep latent representations from the action head, UniReflex drives a fast reflex network that decouples active force exertion from external interaction response. This scheme predicts normalized anisotropic stiffness directions for directional compliance allocation. Furthermore, UniReflex integrates an adaptive gating mechanism that enables seamless transitions between position-dominant planning and force-dominant execution. Real-world bimanual experiments demonstrate that UniReflex significantly improves contact stability and success rates while preserving original position accuracy. Our approach achieves 25-66x lower per-step backward latency relative to joint training strategies on the evaluated backbones.

Problem

• generative imitation learning
• force control
• closed-loop force regulation
• variable impedance control
• trajectory planning

Innovation

• Plug-and-Play Framework
• Variable Impedance Control (VIC)
• Fast Reflex Network
• Adaptive Gating Mechanism