A Learning Framework for Diverse Legged Robot Locomotion Using Barrier-Based Style Rewards

2024-09-24

arXiv.org

AI Summary

This work addresses the challenge of multimodal locomotion control for legged robots by proposing a model-free reinforcement learning framework enabling dynamic switching among quadrupedal, tripedal, and bipedal gaits. Methodologically, it introduces a novel motion-style reward mechanism based on relaxed logarithmic barrier functions, which softly encodes gait periodicity and enables real-time adaptation of gait parameters—including foot clearance height, joint configuration, and body height—facilitating end-to-end policy training without external perception inputs. Key contributions include the first unified realization of autonomous multimodal locomotion on a 45 kg heavy-duty robot (KAIST HOUND/HOUND2): quadrupedal galloping at 4.67 m/s and obstacle negotiation up to 67 cm; bipedal running at 3.6 m/s, carrying a 7.5 kg payload, and autonomous stair climbing; and stable tripedal locomotion. The framework demonstrates robustness, adaptability, and scalability across diverse gait modes under realistic hardware constraints.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This work introduces a model-free reinforcement learning framework that enables various modes of motion (quadruped, tripod, or biped) and diverse tasks for legged robot locomotion. We employ a motion-style reward based on a relaxed logarithmic barrier function as a soft constraint, to bias the learning process toward the desired motion style, such as gait, foot clearance, joint position, or body height. The predefined gait cycle is encoded in a flexible manner, facilitating gait adjustments throughout the learning process. Extensive experiments demonstrate that KAIST HOUND, a 45 kg robotic system, can achieve biped, tripod, and quadruped locomotion using the proposed framework; quadrupedal capabilities include traversing uneven terrain, galloping at 4.67 m/s, and overcoming obstacles up to 58 cm (67 cm for HOUND2); bipedal capabilities include running at 3.6 m/s, carrying a 7.5 kg object, and ascending stairs-all performed without exteroceptive input.

Problem

• Enabling diverse locomotion modes for legged robots
• Using barrier-based rewards to guide motion style learning
• Achieving robust locomotion without exteroceptive input

Innovation

• Model-free reinforcement learning for diverse locomotion
• Barrier-based style rewards for motion constraints
• Flexible gait cycle encoding for adaptive learning