AI Insight
This study introduces PUMA, an end-to-end machine learning framework that enables quadruped robots to perform parkour maneuvers by integrating visual perception with foothold selection in a unified system. Unlike traditional hierarchical approaches that rely on pre-computed foothold positions, PUMA uses terrain features to estimate relative distance and heading information, allowing robots to adaptively select footholds in real-time during obstacle traversal. Testing in both simulated and real-world environments demonstrated improved agility and robustness across various complex discrete terrains.
Why it matters
This advancement could enhance the capabilities of legged robots in challenging environments such as search and rescue operations, industrial inspection of uneven terrain, or exploration missions where adaptive locomotion is critical. By eliminating the need for pre-programmed foothold sequences, the system enables more flexible and autonomous navigation in unpredictable settings.
Understand the Science
arXiv:2601.15995v2 Announce Type: replace-cross
Abstract: Parkour tasks for quadrupeds have emerged as a promising benchmark for agile locomotion. While human athletes can effectively perceive environmental characteristics to select appropriate footholds for obstacle traversal, endowing legged robots with similar perceptual reasoning remains a significant challenge. Existing methods often rely on hierarchical controllers that follow pre-computed footholds, thereby constraining the robot’s real-time adaptability and the exploratory potential of reinforcement learning. To overcome these challenges, we present PUMA, an end-to-end learning framework that integrates visual perception and foothold priors into a single-stage training process. This approach leverages terrain features to estimate egocentric polar foothold priors, composed of relative distance and heading, guiding the robot in active posture adaptation for parkour tasks. Extensive experiments conducted in simulation and real-world environments across various discrete complex terrains, demonstrate PUMA’s exceptional agility and robustness in challenging scenarios.
Source: PUMA: Perception-driven Unified Foothold Prior for Mobility Augmented Quadruped Parkour