|
Associate Professor Department of Mathematics and Computer Science The University of North Carolina at Pembroke
|
 |
Machine Learning (ML) algorithms are increasingly used in the medical field to manage and diagnose diseases, including Covid-19 patients who have a higher incidence of blood clots. Current methods for reporting thrombogenesis-related fluid dynamic metrics for patient-specific anatomies can take weeks to months for a single patient. To address this issue, we propose a ML-based method for rapid thrombogenesis prediction in the carotid artery of Covid-19 patients. The main contribution of the proposed system is the ability to quickly predict thrombogenesis in Covid-19 patients using ML models, which can help in preventive medicine by detecting serious diseases in advance.
Recent innovations in the development of communication devices and control processors have become cheaper than years before. This has provided the opportunity for all devices to have the ability to communicate and share data in the ever growing ecology that makes up the Internet of Things (IoT). We currently have smart light bulbs, smart thermostats, washer and dryer machines, motion sensors, and web cameras connected to the Internet and we can control these devices using smart phone applications and other hand-held devices. Like the examples above, billions of the Internet of Things (IoT) devices around the world are connected to the Internet and are simultaneously collecting and sharing data. For example, the smart watch and fitness tracker such as Fitbit, Apple watch, Samsung gear can reveal the sensitive information such as location information and activities. The fitness tracking company Strava revealed the heat map that show location and activities such as jog, bike and exercise of solider of U.S. military facilities. This means that it is exposed to security risks. But, as IoT devices does not have the high computation power and enough memory resource due to small device like light bulb, smart thermostat and smart watch we cannot adopt current security algorithm directly. Based on this reason, currently we are actively research a decentralized and private IoT Framework using Blockchain
In ubiquitous environment, mobility of mobile node has increased. Increasing mobility of mobile node we have often experience co-channel and adjacent interference with using same channel and adjacent in enclosed space. These interferences are a hot issue affecting performance degradation in wireless ad-hoc network. And NS3 is the most frequently used simulation tool by network researcher. In this research, we conducted an experiment at outside and simulation in NS3 for co-channel interference. As we compared network performance between real network and simulation we did analysis about performance impact at each network stack and how do errors propagate up from the physical layer to TCP when co-channel interference occurs in wireless ad-hoc network. Our work contributes to a more clear understanding for error propagation to the upper layer of network stack when co-channel interference occurs in ad-hoc network and to validate TCP Performance based on Interference modeling of NS3 simulator.
Recently, vehicular ad hoc network (VANET) is receiving lots of attentions as this new networking technology is expected to improve our daily driving experience greatly and will enable a number of emerging applications. It is envisioned that the vehicles in VANETs are armed with a number of advanced technologies such as wireless transceiver, video cameras, etc. This research investigates the potential of the advanced VANET nodes to construct an impromptu surveillance system to surround an area of interest, which can be a city block, such that any suspect of interest leaving the city block can be monitored by a VANET node participating the surveillance system. Such a system can be useful to provide an emergency response system to keep the track of suspects who are leaving the area by walk or by car after committing a crime inside the block, e.g. Boston bombing suspects.