解决视觉语言动作模型在动作分段执行时的突变问题。
SEAM: Smooth Execution of Action-Chunked Motion for Vision-Language-Action Policies
- 利用前一段未执行尾部作为一致性参考，实现无训练推理时平滑衔接。
- 在LIBERO-10上使边界抖动降低28%，分段不连续性减少27%。
- 无需反向传播或重训练，适合部署在实时机器人控制场景。
视觉-语言-动作（VLA）策略在执行固定长度动作分段时可能出现多模态分叉：由于各分段独立生成于高斯潜变量，相邻分段轨迹模式不一致，导致在分段边界产生突变。现有方法需在去噪每一步进行反向传播、依赖拒绝采样或重新训练，分别以计算开销或任务可靠性为代价。我们提出SEAM（Smooth Execution of Action-Chunked Motion），一种无需训练的推理阶段方法，适用于流匹配型VLA。SEAM利用同步执行的简单洞察：机器人执行前缀后，前一动作段的未执行尾部已作为解析的一致性参考。其核心机制速度引导损失引导（VLS）从该尾部推导时间相关目标，并在每次欧拉步后施加闭式修正，无需对策略网络反向传播。在LIBERO-10数据集上，pi_0.5条件下，SEAM将边界抖动降低28%，分段过渡不连续性减少27%，保持基线任务成功率，且去噪循环成本接近无引导基线。
原文摘要 · Abstract (English)
Vision-Language-Action (VLA) policies that execute fixed-length action chunks can exhibit multimodal bifurcation: a cross-chunk inconsistency in which adjacent chunks generated from independent Gaussian latents can converge to incompatible trajectory modes, producing abrupt discontinuities at chunk boundaries. Existing remedies either require backpropagation through the policy at each denoising step, rely on rejection sampling, or require retraining, each trading computational cost or task reliability for smoother transitions. We propose SEAM (Smooth Execution of Action-Chunked Motion), a training-free inference-time method for flow matching VLAs. SEAM exploits a simple synchronous-execution insight: after the robot consumes the executed prefix, the previous chunk's unexecuted tail is already available as an analytic consistency reference. Its core mechanism, Velocity-guided Loss Steering (VLS), derives a time-dependent target from this tail and applies a closed-form correction after each Euler step without backpropagating through the policy network. On LIBERO-10 with pi_0.5, SEAM reduces boundary jerk by 28%, reduces chunk transition discontinuity by 27%, preserves baseline-level task success, and keeps denoising-loop cost near the unguided baseline.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。