当前位置: X-MOL 学术J. Ind. Inf. Integr. › 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.)
VR-enabled engineering consultation chatbot for integrated and intelligent manufacturing services
Journal of Industrial Information Integration ( IF 15.7 ) Pub Date : 2022-02-03 , DOI: 10.1016/j.jii.2022.100331
Amy J.C. Trappey , Charles V. Trappey , Min-Hua Chao , Chun-Ting Wu

Industrial equipment manufacturing, such as electric power transformers, require highly customized design, assembly, installation and maintenance services for customers. Engineering consultation is an essential manufacturing service which requires deep domain knowledge. This research develops and implements the framework and the prototype of a Virtual Reality (VR) enabled intelligent engineering consultation chatbot system with natural language communication capabilities. The consultation chatbot case focuses on the power transformer knowledge domain. The system is constructed using a comprehensive knowledge base of Frequently Asked Questions (FAQs) and answers from a power transformer manufacturer. A total of 490,000 technical articles (with more than 1.1 billion words) from Wikipedia and a comprehensive domain-specific document set are used to train a word embedding model for retrieving questions and answers (QA) from the FAQ knowledge base. Immersive VR technology provides a highly interactive QA chatbot with realistic graphical views for informative engineering counselling. The case example demonstrates the accuracy of the proposed VR-enabled chatbot for transformer QAs exceeds 91%.



中文翻译:

面向集成和智能制造服务的VR工程咨询聊天机器人

电力变压器等工业设备制造需要为客户提供高度定制的设计、组装、安装和维护服务。工程咨询是一项必不可少的制造服务,需要深厚的领域知识。本研究开发并实现了具有自然语言通信能力的支持虚拟现实 (VR) 的智能工程咨询聊天机器人系统的框架和原型。咨询聊天机器人案例侧重于电力变压器知识领域。该系统是使用常见问题 (FAQ) 的综合知识库和电力变压器制造商的答案构建的。共计49万篇技术文章(其中1. 来自维基百科的 10 亿字)和一个全面的特定领域文档集用于训练一个词嵌入模型,用于从常见问题知识库中检索问题和答案(QA)。沉浸式 VR 技术提供了一个高度交互的 QA 聊天机器人,具有逼真的图形视图,用于提供信息工程咨询。该案例展示了针对变压器 QA 提出的支持 VR 的聊天机器人的准确度超过 91%。

更新日期:2022-02-06
down
wechat
bug