SWAP: Symmetric Equivariant World-Model for Agile Robot Parkour

2026-06-18

AI Summary

This work addresses the inefficiency and limited generalization of conventional latent-variable world models in quadrupedal parkour, which redundantly model left-right symmetric interactions and struggle to capture underlying geometric regularities. To overcome these limitations, the study introduces symmetry equivariance as a structural prior, integrating it end-to-end into both the latent-variable world model and the Actor-Critic policy network. This approach significantly enhances model compactness and enables strong zero-shot transfer across diverse environments. Real-world experiments demonstrate unprecedented capabilities, including leaping across a 2.13-meter gap and climbing a 1.63-meter platform—setting a new benchmark for quadrupedal parkour—and exhibit robust generalization to unseen mirrored terrains and varied outdoor scenarios.

Abstract

While latent world models enable the proactive predictions required for extreme parkour, their purely data-driven nature forces them to redundantly encode left-right symmetric interactions as independent patterns. This inflates the learning burden and hinders the capture of geometric regularities, restricting the latent space's efficiency for downstream policies. To address this, we propose SWAP, an end-to-end equivariant symmetric world model. This framework embeds symmetry directly into both the world model and the actor-critic networks. In real-world tests, the robot leaps across a 2.13 m gap and climbs a 1.63 m platform, breaking records for quadruped parkour. Furthermore, the framework exhibits robust geometric generalization to unseen mirrored terrains and exceptional zero-shot transferability across diverse outdoor environments. These results demonstrate that symmetry equivariance is an effective structural prior for pushing the physical boundaries of learned legged locomotion.

Problem

• symmetry
• world model
• equivariance
• robot parkour
• geometric generalization

Innovation

• symmetry equivariance
• world model
• robot parkour
• geometric generalization
• zero-shot transfer