外观
论文arXiv 2609.289601 个月前
脚步中的回声：使用 Gated Memory 学习感知型人形 Parkour
原文链接：Echo in the Steps: Learning Perceptive Humanoid Parkour with Gated Memory（arXiv 2609.28960 · 发布日期：2026-09-24）
Echo in the Steps: Learning Perceptive Humanoid Parkour with Gated Memory
面向高度不连续地形中的人形敏捷越障，提出基于显著性引导时序感知与门控记忆的方法，并在仿真和真实世界验证。
核心要点
- 稀疏落脚点和狭窄支撑区域要求精确选点、有效利用视觉观测，并在快速转换中保持交替落脚。
- 框架仅使用机载深度观测，结合显著性先验与门控记忆，跨帧保留有信息量的深度特征。
- 引入交替损失作为对称性正则，鼓励交替步态模式并提升穿越过程的鲁棒性。
- 大量实验显示，该方法在仿真和真实世界的挑战地形上显著提高成功率与落脚准确性。
验证范围与局限：验证范围为具有有限落脚点的挑战地形，包含仿真与真实世界实验。
英文摘要 (Abstract)
While recent advances in perceptive locomotion have enabled humanoid robots to traverse structured terrains, agile parkour in highly discontinuous environments remains an open challenge. In particular, crossing sparse footholds and narrow support regions requires precise foothold selection, effective use of visual observations, and consistent alternating foot placement during fast transitions. In this paper, we present a perceptive humanoid parkour framework that enables stable traversal across terrains with limited foothold availability using only onboard depth observations. The framework features a saliency-guided temporal perception module that combines a saliency prior with gated memory. It retains informative depth features across frames, enabling reliable foot placement from partial observations. By introducing an alternation loss, our symmetry regularization encourages alternating gait patterns and improves traversal robustness. Extensive experiments show that our method significantly improves success rate and foothold accuracy on challenging terrains in both simulation and the real world.
作者：Ming-ju Lee, Zizhuo Wang, Shaoting Zhu, Haozhe Lou, Hang Zhao, Yiming Li
来源
本页事实对应的公开页面，以原文为准。