PALo: Learning Posture-Aware Locomotion for Quadruped Robots

2025-03-06

AI Summary

Achieving simultaneous linear/angular velocity tracking and real-time body pose regulation (height, pitch, roll) for quadrupedal robots on unstructured terrain (e.g., slopes, rubble) remains challenging. Method: This paper proposes PALo, an end-to-end reinforcement learning framework grounded in a Partially Observable Markov Decision Process (POMDP) formulation and realized via real-time closed-loop control. Contribution/Results: PALo introduces three key innovations: (1) a novel posture-aware joint locomotion control paradigm; (2) an asymmetric Actor-Critic architecture that explicitly models observation discrepancies between simulation and reality, mitigating the sim-to-real transfer bottleneck; and (3) a phased curriculum learning strategy enabling zero-shot deployment on physical hardware. Evaluated in simulation, PALo achieves agile, posture-adaptive locomotion and transfers directly to a real quadrupedal robot, demonstrating stable, robust real-time control across diverse complex terrains without fine-tuning.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

With the rapid development of embodied intelligence, locomotion control of quadruped robots on complex terrains has become a research hotspot. Unlike traditional locomotion control approaches focusing solely on velocity tracking, we pursue to balance the agility and robustness of quadruped robots on diverse and complex terrains. To this end, we propose an end-to-end deep reinforcement learning framework for posture-aware locomotion named PALo, which manages to handle simultaneous linear and angular velocity tracking and real-time adjustments of body height, pitch, and roll angles. In PALo, the locomotion control problem is formulated as a partially observable Markov decision process, and an asymmetric actor-critic architecture is adopted to overcome the sim-to-real challenge. Further, by incorporating customized training curricula, PALo achieves agile posture-aware locomotion control in simulated environments and successfully transfers to real-world settings without fine-tuning, allowing real-time control of the quadruped robot's locomotion and body posture across challenging terrains. Through in-depth experimental analysis, we identify the key components of PALo that contribute to its performance, further validating the effectiveness of the proposed method. The results of this study provide new possibilities for the low-level locomotion control of quadruped robots in higher dimensional command spaces and lay the foundation for future research on upper-level modules for embodied intelligence.

Problem

• Develops posture-aware locomotion for quadruped robots.
• Balances agility and robustness on complex terrains.
• Enables real-time control without fine-tuning.

Innovation

• End-to-end deep reinforcement learning framework
• Asymmetric actor-critic architecture for sim-to-real
• Customized training curricula for agile control