# 探索具身操作中组合泛化的极限

> 原文链接：https://robotics.hellobike.com/papers/2026-10-11-2610-15425
> 频道：papers | 更新日期：2026-10-11 | 标签：arxiv, cs.RO
> 核心摘要：针对长程具身操作技能重组难题，提出HSE、RoboCombine和PRTI，并在RoboCombine上验证有序成功率提升。

# 探索具身操作中组合泛化的极限

::: v-pre
> 原文链接：[Exploring the Limits of Combinatorial Generalization in Embodied Manipulation](https://arxiv.org/abs/2610.15425)（arXiv 2610.15425 · 发布日期：2026-10-11）

**Exploring the Limits of Combinatorial Generalization in Embodied Manipulation**

针对长程具身操作技能重组难题，提出HSE、RoboCombine和PRTI，并在RoboCombine上验证有序成功率提升。

## 核心要点

- 长程操作需无组合示范地重组原子技能，完整指令训练在重组上仅7.7%，显示模板记忆。
- HSE通过规划器、触发器、策略的共享模块接口，诊断技能重组并分解长程失败模式。
- RoboCombine评估无组合轨迹的技能重组，揭示失败集中在子任务交接的起始状态差距。
- PRTI拓宽原子任务起始状态覆盖，在oracle timed HSE下有序成功率达18.16%；代码已开源。

**验证范围与局限**：验证范围为受控研究中的RoboCombine基准与oracle timed HSE设置，未包含组合技能训练轨迹。

## 英文摘要 (Abstract)

In embodied manipulation, Vision Language Action (VLA) models can execute atomic skills, but long-horizon manipulation requires recombining these skills without demonstrations for every skill combination. The full instruction route treats skill chaining as long-horizon imitation. In a controlled study, full instruction training on limited skill combination sequences reaches 40.0% ordered success on covered combinations but only 7.7% on recombination, suggesting memorization of sequence templates rather than robust skill recombination. We introduce Hierarchical Subtask Execution (HSE), a diagnostic framework that exposes skill recombination through a shared modular interface over the planner, trigger, and policy, decomposing long-horizon failures into separable modes. We introduce RoboCombine, a benchmark for skill recombination without combined skill training trajectories, with metrics for ordered success and transition conversion across subtask boundaries. RoboCombine reveals that atomic competence alone does not ensure ordered execution. Failures concentrate at transitions where downstream subtasks launch from handoff states from preceding subtasks rather than canonical atomic starts. We identify this mismatch as the start state gap. To address it, we introduce Pose Robust Task Initialization (PRTI), a lightweight atomic task post-training strategy that broadens start state coverage beyond canonical starts. Under oracle timed HSE on RoboCombine, PRTI reaches 18.16% ordered success, compared with 4.21% under random perturbation and 9.47% under a robot-reset oracle, without combined skill training trajectories. These results show that combinatorial generalization requires handoff robust atomic skills that convert first stage progress into reliable downstream execution, not merely decomposition into known subtasks. Code: https://github.com/AutoLab-SAI-SJTU/RoboCombine.

作者：Xupeng Zhang, Yantai Yang, Chang Guo, Zhaokai Yin, Zhipeng Zhang
:::

## 原始参考来源 (Sources)

- [arXiv 2610.15425：Exploring the Limits of Combinatorial Generalization in Embodied Manipulation](https://arxiv.org/abs/2610.15425)