Research

Paper

TESTING March 06, 2026

Task-Level Decisions to Gait Level Control: A Hierarchical Policy Approach for Quadruped Navigation

Authors

Sijia Li, Haoyu Wang, Shenghai Yuan, Yizhuo Yang, Thien-Minh Nguyen

Abstract

Real-world quadruped navigation is constrained by a scale mismatch between high-level navigation decisions and low-level gait execution, as well as by instabilities under out-of-distribution environmental changes. Such variations challenge sim-to-real transfer and can trigger falls when policies lack explicit interfaces for adaptation. In this paper, we present a hierarchical policy architecture for quadrupedal navigation, termed Task-level Decision to Gait Control (TDGC). A low-level policy, trained with reinforcement learning in simulation, delivers gait-conditioned locomotion and maps task requirements to a compact set of controllable behavior parameters, enabling robust mode generation and smooth switching. A high-level policy makes task-centric decisions from sparse semantic or geometric terrain cues and translates them into low-level targets, forming a traceable decision pipeline without dense maps or high-resolution terrain reconstruction. Different from end-to-end approaches, our architecture provides explicit interfaces for deployment-time tuning, fault diagnosis, and policy refinement. We introduce a structured curriculum with performance-driven progression that expands environmental difficulty and disturbance ranges. Experiments show higher task success rates on mixed terrains and out-of-distribution tests.

Metadata

arXiv ID: 2603.05783
Provider: ARXIV
Primary Category: cs.RO
Published: 2026-03-06
Fetched: 2026-03-09 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.05783v1</id>\n    <title>Task-Level Decisions to Gait Level Control: A Hierarchical Policy Approach for Quadruped Navigation</title>\n    <updated>2026-03-06T00:25:07Z</updated>\n    <link href='https://arxiv.org/abs/2603.05783v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.05783v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Real-world quadruped navigation is constrained by a scale mismatch between high-level navigation decisions and low-level gait execution, as well as by instabilities under out-of-distribution environmental changes. Such variations challenge sim-to-real transfer and can trigger falls when policies lack explicit interfaces for adaptation. In this paper, we present a hierarchical policy architecture for quadrupedal navigation, termed Task-level Decision to Gait Control (TDGC). A low-level policy, trained with reinforcement learning in simulation, delivers gait-conditioned locomotion and maps task requirements to a compact set of controllable behavior parameters, enabling robust mode generation and smooth switching. A high-level policy makes task-centric decisions from sparse semantic or geometric terrain cues and translates them into low-level targets, forming a traceable decision pipeline without dense maps or high-resolution terrain reconstruction. Different from end-to-end approaches, our architecture provides explicit interfaces for deployment-time tuning, fault diagnosis, and policy refinement. We introduce a structured curriculum with performance-driven progression that expands environmental difficulty and disturbance ranges. Experiments show higher task success rates on mixed terrains and out-of-distribution tests.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.RO'/>\n    <published>2026-03-06T00:25:07Z</published>\n    <arxiv:comment>Submitted to IROS 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.RO'/>\n    <author>\n      <name>Sijia Li</name>\n    </author>\n    <author>\n      <name>Haoyu Wang</name>\n    </author>\n    <author>\n      <name>Shenghai Yuan</name>\n    </author>\n    <author>\n      <name>Yizhuo Yang</name>\n    </author>\n    <author>\n      <name>Thien-Minh Nguyen</name>\n    </author>\n  </entry>"
}