VLM-DEWM：动态外部世界模型用于制造中的可验证和抗毁视觉语言规划
VLM-DEWM: Dynamic External World Model for Verifiable and Resilient Vision-Language Planning in Manufacturing
AI总结：
VLM-DEWM通过动态外部世界模型提升制造中视觉语言规划的可验证性和抗毁性，显著提高状态跟踪精度和恢复成功率。
摘要：
视觉语言模型（VLM）在智能制造中的高层规划中显示出潜力，但其在动态工作单元中的部署面临两个关键挑战：（1）无状态操作，无法持续跟踪视线外的状态，导致世界状态漂移；（2）透明推理，故障难以诊断，导致成本高昂的盲目重试。本文提出了VLM-DEWM，一种将VLM推理与世界状态管理解耦的认知架构，通过持久且可查询的动态外部世界模型（DEWM）。每个VLM决策都被结构化为可外部化的推理轨迹（ERT），包括动作提案、世界信念和因果假设，这些在执行前会与DEWM进行验证。当发生故障时，通过预测状态与观测状态之间的差异分析，可以实现针对性恢复，而不是全局重新规划。我们评估了VLM-DEWM在多站装配、大规模设施探索和真实机器人恢复中的表现，针对诱导故障。与基线记忆增强VLM系统相比，VLM-DEWM将状态跟踪准确性从56%提高到93%，将恢复成功率从低于5%提高到95%，并通过结构化内存显著减少了计算开销。这些结果确立了VLM-DEWM作为动态制造环境中长时间跨度机器人操作的可验证和抗毁解决方案。
英文摘要：
Vision-language model (VLM) shows promise for high-level planning in smart manufacturing, yet their deployment in dynamic workcells faces two critical challenges: (1) stateless operation, they cannot persistently track out-of-view states, causing world-state drift; and (2) opaque reasoning, failures are difficult to diagnose, leading to costly blind retries. This paper presents VLM-DEWM, a cognitive architecture that decouples VLM reasoning from world-state management through a persistent, queryable Dynamic External World Model (DEWM). Each VLM decision is structured into an Externalizable Reasoning Trace (ERT), comprising action proposal, world belief, and causal assumption, which is validated against DEWM before execution. When failures occur, discrepancy analysis between predicted and observed states enables targeted recovery instead of global replanning. We evaluate VLM-DEWM on multi-station assembly, large-scale facility exploration, and real-robot recovery under induced failures. Compared to baseline memory-augmented VLM systems, VLM DEWM improves state-tracking accuracy from 56% to 93%, increases recovery success rate from below 5% to 95%, and significantly reduces computational overhead through structured memory. These results establish VLM-DEWM as a verifiable and resilient solution for long-horizon robotic operations in dynamic manufacturing environments.
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。
