A Study on Architecture of Interactive Manufacturing System

交互式制造系统体系结构研究

• 批准号: 11450060
• 负责人: UEDA Kanji
• 金额: $ 6.98万
• 依托单位: KOBE UNIVERSITY
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B).
• 财政年份: 1999
• 资助国家: 日本
• 起止时间: 1999 至 2000
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-11450060/
• 关键词: Interactive Manufacturing; Biological Manufacturing; virtual space; self-organization; reinforcement learning; インタラクティブ環境; 人工現実感

This study can be concluded as following :1. Framework of Interactive Manufacturing SystemsFirstly, the difficulties of problems in the current manufacturing have been classified. To overcome the most difficult class, Class III, Interactive Manufacturing Systems has been proposed. To use virtual space has been proposed in order to realize Interactive Manufacturing, and the relationship between virtual and real spaces has been theoretically concerned.2. Interactive Environment Using Virtual SpaceThe method to connect self-organization simulation with virtual space has been proposed to realize an interactive environment. The effectiveness of the proposed method has been verified, because interactions between artifacts and human in the virtual pace have been able to be observed in the experiments.3. Manufacturing Systems Using Reinforcement LearningTo realize interactions between artifacts and human in the Interactive Manufacturing, the artifacts should be able to adapt to the action by human. In this study, manufacturing system using reinforcement learning, where the objective can be achieved as the result of local interactions between autonomous elements, has been proposed. Because the objectives, minimizing deviation from due date and maximizing throughput, have been achieved in the experiments, the effectiveness of the proposed method has been verified.

本研究的主要结论如下：1.交互式制造系统框架首先对当前制造业中存在的问题进行了分类。为了克服最困难的类，类III，交互式制造系统已被提出。提出了利用虚拟空间实现交互式制造，并从理论上分析了虚拟空间与真实的空间的关系.利用虚拟空间实现交互环境提出了将自组织仿真与虚拟空间相结合的方法，实现了一个交互环境。实验中可以观察到虚拟空间中人与工件之间的交互，验证了该方法的有效性.基于强化学习的制造系统为了实现交互式制造中工件与人之间的交互，工件应该能够适应人的行为。在这项研究中，制造系统使用强化学习，其中的目标可以实现的自治元素之间的局部相互作用的结果，已经提出。实验结果表明，该方法能够实现交货期偏差最小化和吞吐量最大化的目标，验证了该方法的有效性。

项目成果

M.Chang: "A Cooperative/Competitive Co-evolutionary Leaning Approach to Sorting Network Problems"Prof.14th World Cngress of the International Federation of Automatic Control(IFAC'99). 187-192 (1999)

M.Chang：“一种对网络问题进行排序的合作/竞争共同进化学习方法”，国际自动控制联合会第 14 届世界大会 (IFAC99) 教授。

K.Ueda: "Synthesis and Emergence"Proceedings of International Work Shop on Emertgent Syntesis-IWES'99. 7-12 (1999)

K.Ueda：“Synthesis and Emergence”国际紧急合成研讨会论文集-IWES99。

J.Vaario: "Adaptive Real-Time Scheduling with Self-Oganization Method"Proceedings of Inernational Work Shop on Emergent Synthesos-IWES'99. 87-92 (1999)

J.Vaario：“采用自组织方法的自适应实时调度”紧急合成国际研讨会论文集-IWES99。

上田完次: "共進化ダイナミクスおよび共生関係の生成"システム制御情報学会論文誌. 14・8(発表予定). (2001)

Kanji Ueda：“共同进化动力学和共生关系的生成”，系统、控制和信息工程师学会汇刊 14/8（待发表）。

K.Ueda: "Reinforcement Learning Approaches to Biological Manufacturing Sysems"Annals of the CIRP(掲載予定). Vol.49,No.1. (2000)

K.Ueda：“生物制造系统的强化学习方法”CIRP 年鉴（即将出版）第 49 卷，第 1 期（2000 年）。