人形机器人对稀疏三维结构的敏捷感知遍历学习

中文摘要 AI 辅助

稀疏三维结构的遍历要求人形机器人在执行敏捷、精确的全身运动时感知薄的、悬垂的几何结构。我们通过单杠遍历研究该问题，机器人必须跳向结构、通过稀疏的单杠交互遍历结构并安全着陆。针对该任务，我们提出一种基于强化学习的感知控制系统，该系统直接基于头戴式固态激光雷达的观测结果运行。为从稀疏返回数据中提取任务相关的几何信息，策略通过带循环记忆的注意力编码器处理原始激光雷达扫描数据。该策略通过分阶段的师生流水线获得，该流水线结合了用于向上跳跃、摆荡和向下跳跃的特权专家策略。为迁移至硬件，我们对激光雷达噪声、电池电压骤降和执行器热极限进行建模，并为人形机器人配备被动钩式末端执行器以实现稳健的单杠交互。在硬件上，得到的策略在三种单杠配置下的15次试验中完成了全部向上跳跃→摆荡→向下跳跃序列的14次，摆荡速度最高可达0.5 m/s。除摆荡外，相同的感知主干网络还支持单独训练的策略，该策略可在横截面积为2 cm的薄悬垂障碍物下俯身通过。

英文摘要

Traversing sparse 3D structures requires humanoid robots to perceive thin, overhanging geometry while executing agile, accurate whole-body motions. We study this problem through monkey-bar traversal, where the robot must jump to the structure, traverse it through sparse bar interactions, and land safely. For this task, we present a reinforcement-learning-based perceptive control system that operates directly on observations from a head-mounted solid-state lidar. To extract task-relevant geometry from the sparse returns, the policy consumes the raw lidar scan through an attention-based encoder with recurrent memory. This policy is obtained by a phase-scheduled teacher- student pipeline that combines privileged experts for jumping up, brachiating, and jumping down. For transfer to hardware, we model lidar noise, battery-voltage sag, and actuator thermal limits, and equip the humanoid with passive hook end-effectors for robust bar interaction. On hardware, the resulting policy completes the full jump-up->brachiation->jump-down sequence in 14 of 15 trials across three bar configurations and reaches brachiation speeds up to 0.5 m/s. Beyond brachiation, the same perception backbone supports a separately trained policy that ducks beneath thin overhead obstacles with 2 cm cross-sections.

发表机构

• Robotic Systems Lab（机器人系统实验室）
• ETH AI Center（苏黎世联邦理工学院人工智能中心）
• Computer Vision and Geometry Group, ETH Zurich（苏黎世联邦理工学院计算机视觉与几何组）

机构由 AI 辅助整理，请以论文原文为准。