WholeBodyMem 为人形机器人引入记忆机制，突破 VLA 策略只看当下的局限
Memory before Goal: Role of Pre-History Memory for Humanoid Loco-Manipulation
人形机器人转身走动就忘了刚才看到的东西，这篇用 SAM3 加关键帧记忆把真机成功率从 36.7% 拉到 88.3%，做机器人方向值得看看具体做法。
arXiv 论文 WholeBodyMem 提出对象引导的感知记忆，让人形机器人在执行 loco-manipulation 任务时保留目标出现前的历史信息。系统用 SAM3 追踪任务对象，由对象事件因果筛选关键帧，动作专家通过 cross-attention 读取这些帧与对象 token。配套基准 HuMeGo 包含 6 个仿真任务和 6 个 Unitree G1 真机任务。结果显示仿真成功率从 29.1% 提升至 72.6%，真机从 36.7% 提升至 88.3%，SIMPLE 标准任务也从 70.2% 提升到 86.9%。
Memory before Goal: Role of Pre-History Memory for Humanoid Loco-Manipulation
Humanoid loco-manipulation in dynamic scenes is non-Markovian: cues vanish as objects change or are hidden, or leave the view as the robot turns and walks, yet whole-body vision-language-action (VLA) policies see only the present. We present WholeBodyMem, an object-guided perceptual memory that keeps the pre-history of object events seen before the robot acts on its goal. SAM3 tracks the task objects, and their events causally select key frames within an adaptive budget, which the action expert reads by cross-attention together with object tokens encoding appearance, geometry and time. We also contribute HuMeGo, a humanoid memory benchmark with six simulated and six real-world Unitree G1 tasks. Keeping the right frames rather than more frames, WholeBodyMem surpasses memoryless and uniform-memory baselines, raising its base policy's success from 29.1% to 72.6% in simulation and from 36.7% to 88.3% on the real robot, while also improving the standard SIMPLE tasks (70.2% to 86.9%).
