| Autor: |
Yin, Kexin, Qi, Qunfen, Morse, Edward, Shakarji, Craig, Srinivasan, Vijay |
| Zdroj: |
Procedia CIRP; 2024, Vol. 129, p163-168, 6p |
| Abstrakt: |
Smart specification and verification of product geometry for manufacturing rely heavily on robust mathematical and computational support. This paper introduces statistical methods that are tailored to address the challenges facing the next generation of ISO GPS standards that govern geometrical product specification and verification. A key focus of this work is the handling of outliers, which are inevitable in industrial metrology. The paper begins by defining outliers based on established statistical literature, standards, and practices. It then explores the application of these concepts to the context of geometrical measurements in ISO GPS standards, while also considering the unique characteristics of structured multidimensional data under the Euclidean distance metric prevalent in this domain. By employing optimization theories and techniques in a unified formulation, the paper extends existing mean and rank-order statistical tools to present preliminary findings for structured data in multidimensions. Furthermore, it outlines potential future research directions concerning mathematical theories and computational techniques aimed at effectively applying median and rank-order statistics, and their incorporation into the future ISO GPS standards. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
