Kaishu Xia

Date of Award

Summer 2021

Document Type

Open Access Dissertation


Mechanical Engineering

First Advisor

Ramy Harik


A manufacturing paradigm shift from conventional control pyramids to decentralized, service-oriented, and cyber-physical systems (CPSs) is taking place in today’s Industry 4.0 revolution. Generally accepted roles and implementation recipes of cyber systems are expected to be standardized in the future of manufacturing industry. Developing affordable and customizable cyber-physical production system (CPPS) and digital twin implementations infuses new vitality for current Industry 4.0 and Smart Manufacturing initiatives. Specially, Smart Manufacturing systems are currently looking for methods to connect factories to control processes in a more dynamic and open environment by filling the gaps between virtual and physical systems.

The work presented in this dissertation first utilizes industrial digital transformation methods for the automation of robotic manufacturing systems, constructing a simulation-based surrogate system as a digital twin to visually represent manufacturing cells, accurately simulate robot behaviors, promptly predict system faults and adaptively control manipulated variables. Then, a CPS-enabled control architecture is presented that accommodates: intelligent information systems involving domain knowledge, empirical model, and simulation; fast and secured industrial communication networks; cognitive automation by rapid signal analytics and machine learning (ML) based feature extraction; and interoperability between machine and human. A successful semantic integration of process indicators is fundamental to future control autonomy. Hence, a product-centered signature mapping approach to automated digital twinning is further presented featuring a hybrid implementation of smart sensing, signature-based 3D shape feature extractor, and knowledge taxonomy. Furthermore, capabilities of members in the family of Deep Reinforcement Learning (DRL) are explored within the context of manufacturing operational control intelligence. Preliminary training results are presented in this work as a trial to incorporate DRL-based Artificial Intelligence (AI) to industrial control processes.

The results of this dissertation demonstrate a digital thread of autonomous Smart Manufacturing lifecycle that enables complex signal processing, semantic integration, automatic derivation of manufacturing strategies, intelligent scheduling of operations and virtual verification at a system level. The successful integration of currently available industrial platforms not only provides facile environments for process verification and optimization, but also facilitates derived strategies to be readily deployable to physical shop floor.

The dissertation finishes with summary, conclusions, and suggestions for further work.