BT-Agent:结合行为树与多保真度评估的机载机器人智能体部署方案
BT-Agent: Deploying Onboard Robot Agents with Behavior Trees and Multi-Fidelity Evaluation
AI总结:
本文提出BT-Agent架构,结合LLM决策与行为树技能,在四足机器人上用9B模型实现完全机载运行,引入多保真度基准评估,在4B、9B规模下较Act/ReAct基准提升显著,结构化工具调用与失败反馈为关键组件。
摘要:
在移动机器人上部署语言模型智能体,需在有限的计算与内存资源内完成任务推理与响应式执行。本文提出BT-Agent,一种机载架构,将大语言模型(LLM)驱动的任务决策与预定义的行为树(Behavior Tree, BT)技能相结合,这些技能可强制执行执行条件并返回结构化失败反馈。我们在四足机器人上使用9B规模的模型,在五个任务类别中实现了完全机载运行。干预测试显示,本地检查可在不等待其他模型决策的情况下停止正在进行的导航,同时智能体保持活跃以重试或报告约束条件。为高效比较智能体设计,我们引入包含161个任务的多保真度基准,对所有任务进行文本评估,对其中39个任务在Habitat平台上进行评估。文本评估的吞吐量比模拟评估高10.9倍,而模拟评估可揭示感知与运动故障如何改变任务结果。BT-Agent在4B和9B规模下,相较于最强的Act/ReAct基准,取得了最大的成功提升,分别约为10和5个百分点。 ablation研究表明,结构化工具调用与信息丰富的失败反馈是最有用的组件。
英文摘要:
Deploying language-model agents onboard mobile robots requires task reasoning and responsive execution within limited compute and memory. We propose BT-Agent, an onboard architecture combining LLM-driven task decisions with predefined behavior-tree (BT) skills. These skills enforce execution conditions and return structured failure feedback. We demonstrate fully onboard operation with a 9B model on a quadruped across five task categories. Intervention tests show that local checks stop ongoing navigation without waiting for another model decision, while the agent remains active to retry or report constraints. To compare agent designs efficiently, we introduce a 161-task multi-fidelity benchmark, evaluated in text on all tasks and in Habitat on 39. Text supports 10.9$\times$ higher evaluation throughput, while simulation reveals how perception and motion failures change task outcomes. BT-Agent's largest success gains over the strongest Act/ReAct baseline occur at 4B and 9B: approximately 10 and 5 percentage points. Ablations identify structured tool calls and informative failure feedback as the most useful components.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。