Figure公司最新组装的F.03人形机器人已能实现自主行走，从生产线直接步行至总部。其核心突破在于仅依靠机载摄像头感知，无需LiDAR或预先地图，即可完成上下楼梯等复杂导航。完整的运动策略完全通过仿真环境中的端到端强化学习训练而成，并零样本迁移至实体机器人。演示中可见其通过神经网络从摄像头数据推断几何环境的深度感知能力，尽管在尺度稳定性和窗户等区域仍存在轻微抖动与伪影。
Freshly assembled Figure's F.03 humanoid can now walk autonomously from the manufacturing line straight to headquarters.
Navigates stairs using only its onboard camera feeds—no LiDAR, no pre-mapped floors.
The full locomotion policy was trained end-to-end with reinforcement learning entirely in simulation, then transferred zero-shot to the physical robot.
Watch its depth perception in action as it handles stair navigation. The colorful reconstruction is how neural networks infer geometry from cameras, though some jitter in scale and artifacts around windows are visible.
F.03 can now walk up/down stairs purely using it's onboard camera perception Our robots now walk from manufacturing when built to HQ This is trained end-to-end with reinforcement learning in simulation在 X 查看被引用的帖子
来源：Rohan Paul · x.com