Hierarchical Equivariant Policy via Frame Transfer

TL;DR

This paper introduces Hierarchical Equivariant Policy (HEP), a framework that leverages domain symmetries and a novel frame transfer interface to improve hierarchical policy learning for robotic manipulation, achieving state-of-the-art results.

Contribution

The paper proposes a new hierarchical policy framework with a frame transfer interface and domain symmetry integration, enhancing robustness and efficiency in complex manipulation tasks.

Findings

State-of-the-art performance in robotic manipulation

Significant improvements in simulation and real-world tasks

Theoretical proof of system equivariance

Abstract

Recent advances in hierarchical policy learning highlight the advantages of decomposing systems into high-level and low-level agents, enabling efficient long-horizon reasoning and precise fine-grained control. However, the interface between these hierarchy levels remains underexplored, and existing hierarchical methods often ignore domain symmetry, resulting in the need for extensive demonstrations to achieve robust performance. To address these issues, we propose Hierarchical Equivariant Policy (HEP), a novel hierarchical policy framework. We propose a frame transfer interface for hierarchical policy learning, which uses the high-level agent's output as a coordinate frame for the low-level agent, providing a strong inductive bias while retaining flexibility. Additionally, we integrate domain symmetries into both levels and theoretically demonstrate the system's overall equivariance. HEP achieves state-of-the-art performance in complex robotic manipulation tasks, demonstrating significant improvements in both simulation and real-world settings.