HiPolicy: Hierarchical Multi-Frequency Action Chunking for Policy Learning

TL;DR

HiPolicy introduces a hierarchical multi-frequency action chunking framework for robotic imitation learning, balancing long-horizon planning and fine control, with adaptive execution based on uncertainty.

Contribution

It proposes a novel hierarchical multi-frequency approach with entropy-guided execution, improving performance and efficiency in robotic policy learning.

Findings

Consistent performance improvements across simulated and real-world tasks.

Enhanced execution efficiency through adaptive, multi-frequency action planning.

Seamless integration into existing generative policies.

Abstract

Robotic imitation learning faces a fundamental trade-off between modeling long-horizon dependencies and enabling fine-grained closed-loop control. Existing fixed-frequency action chunking approaches struggle to achieve both. Building on this insight, we propose HiPolicy, a hierarchical multi-frequency action chunking framework that jointly predicts action sequences at different frequencies to capture both coarse high-level plans and precise reactive motions. We extract and fuse hierarchical features from history observations aligned to each frequency for multi-frequency chunk generation, and introduce an entropy-guided execution mechanism that adaptively balances long-horizon planning with fine-grained control based on action uncertainty. Experiments on diverse simulated benchmarks and real-world manipulation tasks show that HiPolicy can be seamlessly integrated into existing 2D and 3D generative policies, delivering consistent improvements in performance while significantly enhancing execution efficiency.