当前位置: X-MOL 学术IEEE Trans. Autom. Sci. Eng. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Leveraging Iterative Plan Refinement for Reactive Smart Manufacturing Systems
IEEE Transactions on Automation Science and Engineering ( IF 5.6 ) Pub Date : 2020-09-09 , DOI: 10.1109/tase.2020.3018402
Bernhard Wally , Jiri Vyskocil , Petr Novak , Christian Huemer , Radek Sindelar , Petr Kadera , Alexandra Mazak-Huemer , Manuel Wimmer

Industry 4.0 production systems must support flexibility in various dimensions, such as for the products to be produced, for the production processes to be applied, and for the available machinery. In this article, we present a novel approach to design and control smart manufacturing systems. The approach is reactive, that is responds to unplanned situations and implements an iterative refinement technique, that is, optimizes itself during runtime to better accommodate production goals. For realizing these advances, we present a model-driven methodology and we provide a prototypical implementation of such a production system. In particular, we employ Planning Domain Definition Language (PDDL) as our artificial intelligence environment for automated planning of production processes and combine it with one of the most prominent Industry 4.0 standards for the fundamental production system model: IEC 62264. We show how to plan the assembly of small trucks from available components and how to assign specific production operations to available production resources, including robotic manipulators and transportation system shuttles. Results of the evaluation indicate that the presented approach is feasible and that it is able to significantly strengthen the flexibility of production systems during runtime. Note to Practitioners —Smart production is an umbrella for a number of shifts and initiatives that deal with digitization of manufacturing/production systems and related issues and potentials. In this work, we present an approach for utilizing automated planning for creating production plans. This is in contrast to the traditional approach, where recipes are programmed into the production system ahead-of-time. However, automated planning relies on specific languages and tools that are hard to master by nonexperts, which is a factor that strongly limited the utilization of plan-driven approaches for industrial automation in practice. Thus, we propose to generate planning tasks automatically with model-driven engineering techniques. We are utilizing the industrial standard IEC 62264 for the description of the production system, and the academic standard Planning Domain Definition Language (PDDL) for planning. PDDL is handled completely transparent to the user, that is the user is shielded from its complexity by employing the IEC 62264 model as the sole frontend.

中文翻译:

利用迭代计划细化功能来构建反应式智能制造系统

工业4.0生产系统必须在各个方面都具有灵活性,例如要生产的产品,要应用的生产过程以及可用的机械。在本文中,我们提出了一种设计和控制智能制造系统的新颖方法。该方法是被动的,即响应计划外的情况并实施迭代优化技术,即在运行时进行自我优化以更好地适应生产目标。为了实现这些进步,我们提出了一种模型驱动的方法,并提供了这种生产系统的原型实现。尤其是,我们将规划域定义语言(PDDL)作为我们的人工智能环境,用于生产过程的自动规划,并将其与最著名的行业4结合在一起。基本生产系统模型的0个标准:IEC62264。我们展示了如何根据可用组件来计划小型卡车的组装,以及如何将特定的生产操作分配给可用的生产资源,包括机器人操纵器和运输系统穿梭车。评估结果表明,所提出的方法是可行的,并且能够在运行时显着增强生产系统的灵活性。执业者注意 -智能生产是处理制造/生产系统数字化以及相关问题和潜力的许多转变和计划的保护伞。在这项工作中,我们提出了一种利用自动化计划来创建生产计划的方法。这与传统方法相反,在传统方法中,提前将配方编程到生产系统中。但是,自动化计划依赖于非专家难以掌握的特定语言和工具,这是在实践中严重限制了将计划驱动的方法用于工业自动化的因素。因此,我们建议使用模型驱动的工程技术自动生成计划任务。我们利用工业标准IEC 62264来描述生产系统,以及用于规划的学术标准规划域定义语言(PDDL)。PDDL的处理对用户完全透明,也就是说,通过使用IEC 62264模型作为唯一的前端,可以保护用户免受其复杂性的影响。
更新日期:2020-09-09
down
wechat
bug