arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2505.19086cs.ROcs.AIcs.GR
收藏

MaskedManipulator: 通用全身体姿操控

MaskedManipulator: Versatile Whole-Body Manipulation

NVIDIA Israel(NVIDIA以色列分公司);NVIDIA USA(NVIDIA美国分公司);NVIDIA Canada(NVIDIA加拿大分公司);Simon Fraser University Canada(西蒙·弗雷泽大学加拿大分校);NVIDIA;Simon Fraser University(西蒙·弗雷泽大学)

Chen Tessler, Yifeng Jiang, Erwin Coumans, Zhengyi Luo, Gal Chechik, Xue Bin Peng

AI总结:

MaskedManipulator通过生成控制策略实现通用全身体姿操控,提供直观用户控制和复杂交互行为。

摘要:

我们解决了合成具有物理模拟的人类动作以实现全身物体操控的挑战。与以往专注于详细动作追踪、轨迹跟随或遥控的方法不同,我们的框架使用户能够指定多样化的高层次目标,如目标物体姿态或身体姿态。为此,我们引入了MaskedManipulator,这是一种从大规模人体动作捕捉数据训练的跟踪控制器中提炼出的生成控制策略。这种两阶段学习过程使系统能够执行复杂交互行为,同时提供直观的用户控制,以控制角色和物体的动作。MaskedManipulator生成的目标导向操控行为扩展了交互动画系统的能力,使其超越了任务特定的解决方案。

英文摘要:

We tackle the challenges of synthesizing versatile, physically simulated human motions for full-body object manipulation. Unlike prior methods that are focused on detailed motion tracking, trajectory following, or teleoperation, our framework enables users to specify versatile high-level objectives such as target object poses or body poses. To achieve this, we introduce MaskedManipulator, a generative control policy distilled from a tracking controller trained on large-scale human motion capture data. This two-stage learning process allows the system to perform complex interaction behaviors, while providing intuitive user control over both character and object motions. MaskedManipulator produces goal-directed manipulation behaviors that expand the scope of interactive animation systems beyond task-specific solutions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑