PH.D dissertation defense on “Edge Computing Enabled Wireless Sensor Network for Extreme Environments – A Case Study on Hyperspectral Image Classification”
Department of Computer Engineering / College of Engineering at University of Mosul discussed a PH.D dissertation on “Edge Computing Enabled Wireless Sensor Network for Extreme Environments – A Case Study on Hyperspectral Image Classification” for the student (Huthaifa Rabeea Mohammed), supervised by Prof. Dr. Jasim Mohammed Abduljabbar on Tuesday, Nov 07, 2023.
The research paves the way for the concept of edge computing in extreme environments, allowing for data processing to occur locally at the data source. This approach offers several advantages, including reduced latency, increased reliability, decreased bandwidth consumption, enhanced privacy and security, and improved efficiency.
A comprehensive framework is proposed for implementing edge computing in extreme environments, encompassing the necessary hardware and software components for local data processing and seamless data transmission to the cloud. The effectiveness of this framework is demonstrated through a detailed case study, showcasing its ability to reduce latency, increase reliability, and decrease bandwidth usage.
The dissertation findings reveal a significant 99% reduction in processing time when handling image data at the edge, as opposed to cloud processing. Additionally, overall energy consumption is halved, and bandwidth usage is significantly decreased. For example, with the Salinas dataset, data transmission is reduced from 26 megabytes to just a few kilobytes. Similarly, in the case of the Indian Pines dataset, the reduction is from 5.7 megabytes to a few kilobytes. This remarkable improvement is directly linked to the volume of data processed at the edge. These results emphasize the clear advantage of processing data at the edge, especially when dealing with substantial sensor data.