Novel Learning Methods for High-dimensional Data with Applications in Process Modeling and Monitoring
Author(s)
Wang, Qian
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
High-dimensional (HD) data emerge in various scenarios in the manufacturing area, such as machining, rolling or lithographic processes. HD data include signals/profiles, images and point clouds, and they play an indispensable part in the modeling and monitoring of manufacturing processes. Statistical models based on these data are used in the monitoring, control, and optimization of the underlying systems. The overall objective of my thesis is to advance the development of process modeling and monitoring methods in the literature, with specific concentration on the challenges faced by the new types of HD data emerging in the real-world applications, like unstructured images/point clouds, HD data possessing complex non-linear relationship, and aerial thermography imaging, where the methods have not been well developed. We proposed novel learning algorithms to tackle these challenges.
Sponsor
Date
2023-04-19
Extent
Resource Type
Text
Resource Subtype
Dissertation