外观
论文arXiv 2610.1032221 天前
从数字人交互到基于物理的人形技能：交互生成器的物理扎根后训练
原文链接：From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators（arXiv 2610.10322 · 发布日期：2026-10-07）
From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators
DIGHT将数字人交互生成器与人形跟踪策略协同适配，在仿真中用物理偏好后训练并验证交互执行。
核心要点
- 现有两人形交互生成依赖物理跟踪策略，但跟踪能力有限，降低参考动作利用率。
- DIGHT用固定跟踪器在仿真执行多种文本条件交互候选，并从rollout构建可执行性与交互保真偏好。
- 方法采用物理解耦的diffusion DPO对齐预训练生成器，并用生成动作微调跟踪器以提升兼容性。
- 实验表明该方法提升生成动作物理合理性，使仿真人形交互更可靠且更忠实。
验证范围与局限：验证范围为仿真中的人形交互。
英文摘要 (Abstract)
Recent methods have made promising progress in generating interactions between two humanoids, largely relying on physics-based tracking policies to convert digital reference motions into executable trajectories. However, limited tracking capabilities restrict the range of reference motions that can be successfully executed, reducing data utilization. Moreover, even successful tracking does not guarantee physically plausible responses or faithful realization of the intended interactions. In this paper, we introduce DIGHT, a co-adaptive framework that couples a Digital human Interaction Generator with a Humanoid Tracking policy. Our DIGHT first executes multiple text-conditioned interaction candidates in simulation using a fixed tracker. It then constructs physics-grounded preferences from the resulting rollouts, covering both general executability and interaction fidelity. Rather than collapsing these signals into a single scalar reward for candidate ranking, we align the pretrained generator using physics-decoupled diffusion direct preference optimization (DPO), preserving criterion-specific supervision without differentiating through the simulator. To improve executability, preference pairs are derived from tracking error, friction, and floating. Additionally, to improve interaction fidelity, we propose to incorporate force feedback from simulator as a measure of contact fidelity and construct preferences over contact occurrence, location, duration, and force magnitude. The aligned generator then supplies reference motions for fine-tuning the tracker, improving compatibility between generation and physical execution. Extensive experiments demonstrate that our approach not only improves the physical plausibility of generated motions but also enables more reliable and faithful humanoid interactions in simulation.
作者：Kerui Chen, J. Zhang, Kai Lv, Hehe Fan
来源
本页事实对应的公开页面，以原文为准。