Engineering Commons^™

Machine Learning For Multiclass Classification And Prediction Of Alzheimer's Disease, Solale Tabarestani

FIU Electronic Theses and Dissertations

Alzheimer's disease (AD) is an irreversible neurodegenerative disorder and a common form of dementia. This research aims to develop machine learning algorithms that diagnose and predict the progression of AD from multimodal heterogonous biomarkers with a focus placed on the early diagnosis. To meet this goal, several machine learning-based methods with their unique characteristics for feature extraction and automated classification, prediction, and visualization have been developed to discern subtle progression trends and predict the trajectory of disease progression.

The methodology envisioned aims to enhance both the multiclass classification accuracy and prediction outcomes by effectively modeling the interplay between the multimodal …

Novel Attacks And Defenses For Enterprise Internet-Of-Things (E-Iot) Systems, Luis C. Puche Rondon

FIU Electronic Theses and Dissertations

This doctoral dissertation expands upon the field of Enterprise Internet-of-Things (E-IoT) systems, one of the most ubiquitous and under-researched fields of smart systems. E-IoT systems are specialty smart systems designed for sophisticated automation applications (e.g., multimedia control, security, lighting control). E-IoT systems are often closed source, costly, require certified installers, and are more robust for their specific applications. This dissertation begins with an analysis of the current E-IoT threat landscape and introduces three novel attacks and defenses under-studied software and protocols heavily linked to E-IoT systems. For each layer, we review the literature for the threats, attacks, and countermeasures. Based …

Advance Solutions For Voltage Control And Optimization In Smart Grids, Temitayo Olayemi Olowu

FIU Electronic Theses and Dissertations

Voltage optimization and control on distribution networks is of uttermost importance to distribution system operators. The performance of the entire distribution network depends on the voltage profile of the system. The aim of this dissertation is to develop threefold advanced solutions that can effectively address and mitigate the voltage control and optimization issues of the smart grid. The first objective of this thesis is to develop accurate methods to quantify and analyze the expected variability individual and aggregated PV systems which significantly impact the voltage profile of grid-interfaced PV systems. We develop an improved statistical formulation, propose the use power …

Textile-Integrated Wearable Radio Frequency (Rf) Wireless Power Transfer And Harvesting For Battery-Free Medical Sensing, Dieff Vital

FIU Electronic Theses and Dissertations

Body-worn sensing is becoming increasingly important for patient monitoring, particularly the elderly. The vital signs can be directly transferred to their physicians, personal caregivers or used for self-diagnosis. To achieve this, wireless, wearable, and self-powered sensors are required. Wireless RF powering/charging, implemented into clothing, upholstery, and smart dressing presents itself as a convenient and practical solution for powering wearable sensors. In this dissertation, the implementation of such wireless charging platforms that are integrated into clothing, and the integration of medical sensor electronics is shown. Fabric-based wireless charging allows for lightweight, scalable, flexible, battery- less, and comfortable wearing. It is also …