![[2602.15549] VLM-DEWM: Dynamic External World Model for Verifiable and Resilient Vision-Language Planning in Manufacturing](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.15549] VLM-DEWM: Dynamic External World Model for Verifiable and Resilient Vision-Language Planning in Manufacturing
Summary
The paper introduces VLM-DEWM, a novel cognitive architecture designed to enhance vision-language planning in manufacturing by addressing challenges in state tracking and reasoning transparency.
Why It Matters
As manufacturing environments become increasingly complex, effective planning and execution are critical. VLM-DEWM offers a solution that improves state-tracking accuracy and recovery success rates, making it a significant advancement in robotics and AI applications in manufacturing.
Key Takeaways
- VLM-DEWM decouples reasoning from world-state management for better planning.
- It improves state-tracking accuracy from 56% to 93% in dynamic environments.
- The architecture enhances recovery success rates from below 5% to 95% during failures.
- Structured memory reduces computational overhead, making it efficient.
- The approach is applicable to multi-station assembly and real-robot recovery scenarios.
Computer Science > Robotics arXiv:2602.15549 (cs) [Submitted on 17 Feb 2026] Title:VLM-DEWM: Dynamic External World Model for Verifiable and Resilient Vision-Language Planning in Manufacturing Authors:Guoqin Tang, Qingxuan Jia, Gang Chen, Tong Li, Zeyuan Huang, Zihang Lv, Ning Ji View a PDF of the paper titled VLM-DEWM: Dynamic External World Model for Verifiable and Resilient Vision-Language Planning in Manufacturing, by Guoqin Tang and 6 other authors View PDF HTML (experimental) Abstract: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. ...
