• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2022, Vol. 58 ›› Issue (10): 374-382.doi: 10.3901/JME.2022.10.374

• 交叉与前沿 • 上一篇    下一篇

扫码分享

基于情境的CAD模型手势操控技术

龙辉, 郝佳, 牛红伟, 王国新, 张非凡   

  1. 北京理工大学机械与车辆学院 北京 100081
  • 收稿日期:2021-05-05 修回日期:2021-10-15 出版日期:2022-05-20 发布日期:2022-07-07
  • 通讯作者: 郝佳(通信作者),男,1984年出生,博士,副教授,博士研究生导师。主要研究方向为知识工程、智能优化设计与创新设计。E-mail:haojia632@bit.edu.cn
  • 作者简介:龙辉,男,1995年出生,博士研究生。主要研究方向为智能人机交互。E-mail:3120205239@bit.edu.cn;牛红伟,男,1992年出生,博士研究生。主要研究方向为多模态人机交互、智能设计。E-mail:nhwfend@bit.edu.cn
  • 基金资助:
    国家自然科学基金(51505032)和国家部委(JCKY2018204B053)资助项目。

Gesture Control Technology of CAD Model Based on Situation Model

LONG Hui, HAO Jia, NIU Hongwei, WANG Guoxin, ZHANG Feifan   

  1. School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081
  • Received:2021-05-05 Revised:2021-10-15 Online:2022-05-20 Published:2022-07-07

摘要: 自然的设计意图表达是概念设计阶段的重要需求之一,探索自然、直观的交互模式逐渐成为CAD技术的新发展方向。目前,基于手势的自然交互模式已成为实现新型CAD交互模式的重要途径。然而,现有基于手势的CAD交互方法中仍存在着映射关系固定,难以利用少量手势表达较多的操控指令的瓶颈问题。为此,提出一种基于情境的CAD模型手势操控技术,建立一种描述CAD建模过程中人机交互场景的情境模型,通过试验采集用户在CAD建模情境中的手势偏好,利用决策树模型建立在特定情境下手势与CAD指令的映射关系,实现手势与CAD指令一对多柔性映射,手势可映射CAD指令数量增加100%。进一步通过试验对比手势和键盘鼠标两种交互模式在CAD建模过程中的操控速度和精确度,结果显示手势交互模式能够更加快速地表达出用户设计意图,且操控精度与键盘鼠标交互模式几乎无差异。

关键词: 手势识别, 情境模型, 概念设计, 智能交互, 计算机辅助设计

Abstract: The natural expression of design intent is one of the important requirements in the conceptual design stage. The exploration of natural and intuitive interaction modes has gradually become the new development direction of computer aided design (CAD) technology. At present, the natural interaction mode based on gestures has become an important way to realize the new CAD interaction mode. However, the existing gestures and CAD interaction modes still have the problems of fixed mapping relationship and insufficient scalability, and it is difficult to use a few gestures to express more control commands. Therefore, a context-based CAD model gesture control technology is proposed. A situation model describing human-computer interaction scene in CAD modeling process is established. The gesture preferences of users in CAD modeling situations were collected through experiments, and the mapping relationship between gesture and CAD instructions was established by decision tree model in specific situations. The one-to-multi-flexible mapping between gesture and CAD instructions was realized, and the number of CAD instructions mapped by gesture increased by 100%. Furthermore, the control speed and accuracy of gesture and keyboard and mouse interaction modes in the process of CAD modeling are compared through experiments. The results show that the gesture interaction model can express the user’s design intention more quickly, and the control accuracy has no difference with the keyboard and mouse interaction model.

Key words: gesture recognition, situation model, conceptual design, intelligent interaction, computer-aided design

中图分类号: