Perceptive Humanoid Parkour: Chaining Dynamic Human Skills via Motion Matching

TL;DR

This paper introduces Perceptive Humanoid Parkour (PHP), a modular framework enabling humanoid robots to perform dynamic, vision-based parkour with long-horizon skill composition, perception-driven decision-making, and real-world adaptability.

Contribution

The paper presents a novel framework combining motion matching, reinforcement learning, and perception for autonomous, dynamic humanoid parkour in complex environments.

Findings

Successfully performed long-horizon obstacle traversal in real-world tests.

Climbed obstacles up to 1.25 meters high, nearly the robot's full height.

Demonstrated real-time, perception-based decision-making for obstacle negotiation.

Abstract

While recent advances in humanoid locomotion have achieved stable walking on varied terrains, capturing the agility and adaptivity of highly dynamic human motions remains an open challenge. In particular, agile parkour in complex environments demands not only low-level robustness, but also human-like motion expressiveness, long-horizon skill composition, and perception-driven decision-making. In this paper, we present Perceptive Humanoid Parkour (PHP), a modular framework that enables humanoid robots to autonomously perform long-horizon, vision-based parkour across challenging obstacle courses. Our approach first leverages motion matching, formulated as nearest-neighbor search in a feature space, to compose retargeted atomic human skills into long-horizon kinematic trajectories. This framework enables the flexible composition and smooth transition of complex skill chains while preserving the elegance and fluidity of dynamic human motions. Next, we train motion-tracking reinforcement learning (RL) expert policies for these composed motions, and distill them into a single depth-based, multi-skill student policy, using a combination of DAgger and RL. Crucially, the combination of perception and skill composition enables autonomous, context-aware decision-making: using only onboard depth sensing and a discrete 2D velocity command, the robot selects and executes whether to step over, climb onto, vault or roll off obstacles of varying geometries and heights. We validate our framework with extensive real-world experiments on a Unitree G1 humanoid robot, demonstrating highly dynamic parkour skills such as climbing tall obstacles up to 1.25m (96% robot height), as well as long-horizon multi-obstacle traversal with closed-loop adaptation to real-time obstacle perturbations.