# SharedKV-BT：行为树智能体的节点局部类型化决策

> 原文链接：https://robotics.hellobike.com/papers/2026-10-05-2610-07327
> 频道：papers | 更新日期：2026-10-05 | 标签：arxiv, cs.RO
> 核心摘要：SharedKV-BT用行为树节点局部字段与Shared-KV并行评分，缓解互依决策延迟，并在操作、导航和电脑使用任务验证。

# SharedKV-BT：行为树智能体的节点局部类型化决策

::: v-pre
> 原文链接：[SharedKV-BT: Node-Local Typed Decisions for Behavior-Tree Agents](https://arxiv.org/abs/2610.07327)（arXiv 2610.07327 · 发布日期：2026-10-05）

**SharedKV-BT: Node-Local Typed Decisions for Behavior-Tree Agents**

SharedKV-BT用行为树节点局部字段与Shared-KV并行评分，缓解互依决策延迟，并在操作、导航和电脑使用任务验证。

## 核心要点

- 自回归决策接口灵活但逐token生成有延迟，现有共享前缀方法未建模依赖或验证执行。
- 每个活跃BT节点暴露阶段局部字段和候选，Shared-KV并行评分后交给独立执行系统。
- 三类任务中，类型化决策比匹配提示的自回归解码快2.36至4.15倍。
- 操作任务上，联合决策准确率由75%升至94%，闭环成功率由0%升至60%。

**验证范围与局限**：验证范围为摘要所述的机器人操作、移动导航和计算机使用任务，未说明仿真或真机设置。

## 英文摘要 (Abstract)

Agent tasks require sequences of interdependent decisions. Autoregressive models support more flexible decision interfaces than conventional classifiers but incur the latency of token-by-token generation. Recent shared-prefix methods reduce this cost by reusing encoded context and scoring multiple decisions in parallel, but do not model decision dependencies or verify execution. We propose SharedKV-BT, where each active node of a behavior tree (BT) exposes stage-local fields and candidates, and Shared-KV scores the candidates in parallel and passes the selected decision to a separate execution system. We tested SharedKV-BT on robot manipulation, mobile navigation, and computer-use tasks. Across three tasks, SharedKV-BT made typed decisions 2.36-4.15 times faster than prompt-matched autoregressive decoding. On the manipulation task, node-local Shared-KV improved joint decision accuracy from 75% to 94% and closed-loop success from 0% to 60%. Fixed-score policy replay showed that stage gating prevented out-of-order actions and external postconditions prevented premature completion.

作者：Naoki Wake, Justin Wagle
:::

## 原始参考来源 (Sources)

- [arXiv 2610.07327：SharedKV-BT: Node-Local Typed Decisions for Behavior-Tree Agents](https://arxiv.org/abs/2610.07327)