Skill-Nav: Enhanced Navigation with Versatile Quadrupedal Locomotion via Waypoint Interface

2025-06-26

AI Summary

To address the insufficient coordination between navigation and complex gait skills in quadrupedal robots, this paper proposes a hierarchical navigation framework based on a waypoint interface. The framework introduces waypoints as a unified semantic interface—novelly decoupling high-level task planning (e.g., LLM-based or classical path planning) from low-level gait control—enabling terrain-adaptive gait generation and real-time obstacle avoidance. Our method integrates deep reinforcement learning, waypoint-guided gait policies, and sim-to-real transfer training. Evaluations on both simulation and physical quadruped platforms demonstrate that, compared to conventional velocity-command interfaces, our approach improves long-range navigation success rate on unstructured terrain by 42%, triples the number of generalizable terrain types, and significantly enhances cross-terrain navigation flexibility and robustness.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Quadrupedal robots have demonstrated exceptional locomotion capabilities through Reinforcement Learning (RL), including extreme parkour maneuvers. However, integrating locomotion skills with navigation in quadrupedal robots has not been fully investigated, which holds promise for enhancing long-distance movement capabilities. In this paper, we propose Skill-Nav, a method that incorporates quadrupedal locomotion skills into a hierarchical navigation framework using waypoints as an interface. Specifically, we train a waypoint-guided locomotion policy using deep RL, enabling the robot to autonomously adjust its locomotion skills to reach targeted positions while avoiding obstacles. Compared with direct velocity commands, waypoints offer a simpler yet more flexible interface for high-level planning and low-level control. Utilizing waypoints as the interface allows for the application of various general planning tools, such as large language models (LLMs) and path planning algorithms, to guide our locomotion policy in traversing terrains with diverse obstacles. Extensive experiments conducted in both simulated and real-world scenarios demonstrate that Skill-Nav can effectively traverse complex terrains and complete challenging navigation tasks.

Problem

• Integrating quadrupedal locomotion skills with navigation for long-distance movement
• Developing a waypoint-guided locomotion policy for obstacle avoidance
• Enabling versatile navigation using general planning tools like LLMs

Innovation

• Waypoint-guided locomotion policy via deep RL
• Hierarchical navigation framework with waypoints
• Integration of LLMs and path planning algorithms