DynamicWAM：面向动态操作的世界-动作模型的双路径运动条件机制
DynamicWAM: Dual-Path Motion Conditioning for World-Action Models in Dynamic Manipulation
AI总结：
针对现有世界-动作模型在动态场景中运动感知与响应控制不足的问题，提出采用双路径运动条件机制的DynamicWAM，在DOMINO数据集及12项真实世界任务中均显著优于基线方法。
AI中文摘要：
动态操作要求机器人推断目标运动并及时响应，但现有的世界-动作模型（World-Action Models, WAMs）通常仅基于当前帧进行条件设置，并同步运行大型主干网络，限制了其在动态场景中的运动感知与响应式控制能力。我们提出DynamicWAM，这是一种用于动态物体操作的紧凑型WAM，采用双路径运动条件机制。DynamicWAM引入历史流条件机制，通过冻结的预训练视频变分自动编码器（video VAE）编码时间对齐的光流帧以及当前观测结果，以保留空间运动结构；同时将位移、持续时间、速度和加速度的运动学描述符注入动作专家模块，以提供运动幅度与时间信息。两条互补路径通过联合世界-动作注意力机制进行融合。此外，通过蒸馏得到的紧凑型主干网络以及基于实时分块（real-time chunking, RTC）的异步执行机制，进一步实现了响应式控制。在DOMINO数据集上，DynamicWAM达到了38.2%的成功率和53.2的操作得分，优于所有评估的基线方法；在涵盖线性、圆形和复合目标运动的12项真实世界任务中，其平均成功率达到46.7%，比最强基线高出22.9个百分点。
英文摘要：
Dynamic manipulation requires robots to infer target motion and respond promptly, yet existing World-Action Models (WAMs) typically condition only on the current frame and execute large backbones synchronously, limiting motion awareness and responsive control in dynamic scenes. We propose DynamicWAM, a compact WAM for dynamic object manipulation with dual-path motion conditioning. DynamicWAM introduces history-flow conditioning, encoding temporally aligned optical-flow frames alongside the current observation through a frozen pretrained video VAE to preserve spatial motion structure, while injecting kinematic descriptors of displacement, duration, velocity, and acceleration into the action expert to provide motion magnitude and timing. The two complementary paths are fused through joint world-action attention. A distilled compact backbone and real-time chunking (RTC)-based asynchronous execution further enable responsive control. On DOMINO, DynamicWAM achieves a 38.2% success rate and a 53.2 manipulation score, outperforming all evaluated baselines. Across 12 real-world tasks spanning linear, circular, and compound target motion, it achieves a 46.7% average success rate, exceeding the strongest baseline by 22.9 percentage points.
