NVIDIA 发布 Metatron:用目标接口让编码智能体驱动人形机器人全身运动
carlosdponx · x · 2026-10-29
NVIDIA researchers unveiled Metatron, a promptable whole-body execution layer (a System 1 controller) for humanoids: a goal-based interface connects System 2 planners—including coding agents—to whole-body movement.
Key ideas:
- Agents specify "what and where": hand/foot/torso targets, travel velocity and heading, arrival times—constraining only what matters while the policy generates the rest
- One frozen G1 policy handles six behaviors: steering, teleoperation, sit & stand, recovery, physical guidance (free arms), and dense whole-body tracking
- Coding agents can write and refine goals to build new skills without retraining the policy
The team showed capable simulated climbing and sliding maneuvers. Project page, paper, and code are public; collaborators include Chen Tessler, Yifeng Jiang, and Xue Bin Peng.
所属事件:NVIDIA 发布 Metatron，让编码智能体驱动人形机器人全身运动(2 条相关)→
「编程与Agent」频道最新
- JobNimbus AI 负责人直播演示：把 Claude Code skill 升级为自改进循环 — aakashgupta · 2026-10-29
- GitHub Copilot Projects 实测:用一个 Agent 管理所有 Agent 会话 — 0xkarasy · 2026-10-29
- AI 编码交接的「验证税」:重建上下文才是隐性成本 — tallmetommy · 2026-10-29
- Greptile TREX 用 Computer Use 端到端测试桌面、Web 与 CLI — ycombinator · 2026-10-29
- K-Dense 开源 177 项科学技能,把 Claude Code 变成科研助手 — alvelda · 2026-10-29
- 纳德拉发布 Copilot Fusion：快 40%、成本降 77%，重构智能体 SDLC — satyanadella · 2026-10-29