计算机集成制造系统 ›› 2015, Vol. 21 ›› Issue (第4期): 1142-1149.DOI: 10.13196/j.cims.2015.04.029

• 产品创新开发技术 • 上一篇    下一篇

基于D-S证据理论的产品多信息融合标识失效补救方法

王健,何卫平,王伟,李夏霜,郭改放   

  1. 西北工业大学现代设计与集成制造教育部重点实验室
  • 出版日期:2015-04-30 发布日期:2015-04-30
  • 基金资助:
    国家自然科学基金资助项目(51275419);国防基础科研资助项目(A2720110011)。

Remediation of failed identification in product multi-information fusion based on D-S evidence theory

  • Online:2015-04-30 Published:2015-04-30
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.51275419),and the National Defense Basic Research Program,China(No.A2720110011).

摘要: 针对离散制造企业生产环境和工艺的复杂性容易引起Data Matrix二维符号标记的磨损、污染和腐蚀而发生标识失效的问题,提出一种基于D-S证据理论的产品多信息融合标识失效补救方法。分析了产品物理与制造过程中多源信息之间的关联和数据结构,建立了离散制造产品零散与不完整多源信息的数学模型;设计了改进的变异系数加权法,并采用基于类中心的变权欧式距离法对失效标识产品特征和历史数据库进行相似性测度计算;通过D-S证据理论对失效标识的产品特征层进行融合辨识。实验结果表明,该方法能够较好地恢复补救产品失效的标识。

关键词: 直接标刻技术, Data Matrix, 恢复补救, 多信息融合, D-S证据理论

Abstract: Aiming at the problem that the two-dimensional Data Matrix symbol was wear,pollution and corrosion to fail identify leaded by production environment and process complexity in discrete manufacturing enterprise,a multi-information fusion to remediation failure identification based on D-S evidence theory was put forward.The correlation and data structure between multi-sources information in product physical and manufacturing process was analyzed,and a mathematical model of discrete manufacturing products'scattered and incomplete multi-sources information was established.The improved weight coefficient of variation was designed.The failure identification of product characterization and the historical database was calculated similarly based on class center Euclidean distance method.The product features layers of failure identification were fused identified by D-S evidence theory.The experimental results demonstrated that the failure identification of product could be better recovered.

Key words: direct marking techniques, Data Matrix, recovery or remediation, multi-information fusion, D-S evidence theory

中图分类号: