FocalPolicy: Frequency-Optimized Chunking and Locally Anchored Flow Matching for Coherent Visuomotor Policy

ICML 2026

Qian He1,2,*, Zhenshuo Yang1,2,*, Wenqi Liang3, Chunhui Hao1, Nicu Sebe3, Jiandong Tian1
1State Key Laboratory of Robotics and Intelligent Systems, Shenyang Institute of Automation, CAS
2University of the Chinese Academy of Sciences
3University of Trento
*Equal contribution
FocalPolicy Teaser

FocalPolicy employs a Foresight Composite Objective to synergize proximal precision with distal coherence across chunks.

Abstract

Visuomotor policies aim to learn complex manipulation tasks from expert demonstrations. However, generating smooth and coherent trajectories remains challenging, as it requires balancing proximal precision with distal foresight. Existing approaches typically focus on optimizing intra-chunk action distributions, often neglecting the inter-chunk coherence. Consequently, inter-chunk discontinuities significantly impede the learning of coherent long-horizon actions.

To overcome this limitation and achieve a synergetic balance between precision and foresight, we propose FocalPolicy, a foresight-aware visuomotor policy that combines Frequency-Optimized Chunking with Locally Anchored flow matching. We introduce a foresight composite objective that supervises time-domain alignment within the proximal actions while regularizing frequency-domain structure over multiple future action chunks to improve cross-chunk coherence. To efficiently learn complex action distributions, we design locally anchored sampling to enhance target signal propagation efficiency during consistency flow matching training. Extensive experiments demonstrate that our method consistently outperforms existing approaches.

Methodology

FocalPolicy Pipeline

The pipeline of FocalPolicy. We propose Locally Anchored Sampling (LAS) to improve the training efficiency of consistency flow matching. The policy is optimized via a Foresight Composite Objective (FCO), which synergizes proximal precision with distal coherence.

Real-World Execution

Demonstrating stable long-horizon execution on multi-stage tasks like Tower Stacking and Cup Matching.

Real World Experiments

BibTeX

@inproceedings{he2026focalpolicy,
  title={FocalPolicy: Frequency-Optimized Chunking and Locally Anchored Flow Matching for Coherent Visuomotor Policy},
  author={He, Qian and Yang, Zhenshuo and Liang, Wenqi and Hao, Chunhui and Sebe, Nicu and Tian, Jiandong},
  booktitle={Proceedings of the 43rd International Conference on Machine Learning},
  year={2026}
}