Date of Award

2026-08-01

Degree Name

Doctor of Philosophy

Department

Computer Science

Advisor(s)

Deepak K. Tosh

Abstract

The advent of Industry 5.0 envisions smart manufacturing characterized by human centricity, sustainability, and systemic resilience. Realizing this vision requires the seamless convergence of Information Technology (IT) and Operational Technology (OT) networks. However, integrating massive, stochastic IT edge computing workloads with deterministic physical control loops introduces severe architectural friction, inherently threatening the safety guarantees required by industrial machinery. To resolve this fundamental incompatibility, this dissertation proposes the Edge-Augmented Real-Time Industrial Control System (EA-RICS).

EA-RICS is a comprehensive, multi-layered architecture designed to dismantle systemic bottlenecks across the physical data plane, the centralized control plane, and the edge operating system. First, the framework leverages P4 programmable data planes to enable autonomous, localized congestion handling, eradicating the latency penalties of centralized SDN controllers. Second, it establishes a mixed criticality traffic engineering methodology, strategically combining Time-Aware, Asynchronous, and Credit-Based shapers to guarantee temporal isolation on shared Ethernet mediums. Third, to efficiently scale across redundant mesh topologies, the architecture integrates exact Pay-Burst-Only-Once (PBOO) network calculus into a lazy-evaluating multipath routing algorithm, maximizing flow admissibility without mathematical overestimation.

Transitioning to the IT edge, the framework strictly bounds the execution latency of visual anomaly detection models through a joint optimization of Once-For-All (OFA) neural architectures and greedy coreset subsampling, successfully resolving Feature Blindness while satisfying hard operational deadlines. Finally, to safely close the industrial control loop, an Extended Berkeley Packet Filter (eBPF) and eXpress Data Path (XDP) ingress gateway bypasses the standard operating system network stack, transforming best effort packet delivery into statistical determinism. Extensive empirical evaluations demonstrate that the EA-RICS framework harmonizes high fidelity edge intelligence with absolute physical determinism, providing a highly scalable and resilient blueprint for next generation industrial automation.

Language

en

Provenance

Received from ProQuest

File Size

203 p.

File Format

application/pdf

Rights Holder

Taposh Kumer Sarker

Download

Share

COinS