论文arXiv 2611.221324 天前
DailyBrick:多阶段智能体系统生成乐高组装指令赋能机器人操作
本文由 AI 根据公开内容整理,可能有误,请以原文为准:DailyBrick: Language-Guided Synthesis of Brick Assembly Instructions(arXiv 2611.22132)
核心内容
针对乐高模型生成需满足物理支撑与无遮挡约束的难题,arXiv论文提出DailyBrick多阶段智能体系统。该方法摒弃训练专家大模型的传统路径,将长上下文连接问题转化为类人程序编写任务,确保每步中间状态均可执行。系统不仅生成3D模型,还同步产出供人类与机器人使用的逐块组装指令,在简单用户请求下的生成质量显著优于近期同类智能体系统。
技术架构采用多阶段Harness辅助LLM生成,核心包含版本化构建SDK与外部Meta-Reviewer智能体。SDK将语义子部件到单块积木的多级决策编译为标准操作;Meta-Reviewer通过重放检查生成包,验证模型及构建过程,定位失败阶段并路由至有界修复或工具更新候选。实验证实该框架有效支持下游机器人基于指令和积木集的自动化装配操作。
关键看点
- 程序化生成范式:将乐高生成重构为可执行程序编写,替代专家LLM训练,确保中间步骤物理可执行性
- 版本化构建SDK:编译从语义子部件到个体积木的多级决策,将复杂结构转化为标准化原子操作序列
- Meta-Reviewer验证:外部智能体重放检查生成包,自动追溯失败阶段并触发有界修复或技能库更新
- 机器人装配验证:生成逐块组装指令直接服务于机械臂操作,在模型装配任务中性能超越现有智能体
工程与产业观察
DailyBrick展示了将具身任务规划从端到端生成转向“代码化+验证闭环”的工程价值。其版本化SDK与Meta-Reviewer机制为解决长序列操作中的累积误差提供了通用范式,对复杂装配任务的Sim2Real迁移及技能库自动化迭代具有重要参考意义。
英文摘要 (Abstract)
Generating a LEGO model from a user request is more than a plausible 3D shape generation. LEGO models require physical support per brick, and each brick should be placed unblocked. Current methods address this issue by training an expert large language model. We reduce the problem of generating extremely long context on all brick connections to that of writing a program that builds LEGO models like a human, considering each intermediate step could be executed. We introduce DailyBrick , a multi-stage agentic system that produces a LEGO model together with a brick-by-brick assembly instruction for human building and robot assembly. We design a multi-stage harness with roles to help LLMs generate more effectively. A versioned construction SDK compiles multi-level decisions, from semantic subparts to individual bricks, into canonical operations. An external Meta-Reviewer agent replays and inspects the generated package, verifies the model and its construction process, identifies failures, traces them to the responsible stage, and routes them to bounded repairs or to editable candidates for later skill or tool updates. Extensive experiments demonstrate that DailyBrick generates LEGO models based on simple user requests better than recent agentic systems, and show its utility in downstream applications such as robot manipulations on model assembly from instructions and brick sets (https://SpatialAgent.github.io/DailyBrick/).
作者:Hang Zhang, Qijian Tian, Yuan Xie, Xuhong Wang, Xin Tan
外部来源
本页事实对应的公开页面,以原文为准。(点击将离开本站)