Optimizing Engineering Tools Using Modern Ground Architectures
Abstract
Over the past decade, a deluge of large and complex datasets (aka big data) has overwhelmed the scientific community. Traditional computing architectures were not capable of processing the data efficiently, or in some cases, could not process the data at all. Industry was forced to reexamine the existing data processing paradigm and develop innovative solutions to address the challenges. This thesis investigates how these modern computing architectures could be leveraged by industry and academia to improve the performance and capabilities of engineering tools. First, the effectiveness of MathWorks Parallel Computing Toolkit is assessed when performing somewhat basic computations in MATLAB. Next, a more computationally intensive series of tests using synthetic aperture radar datasets is demonstrated using the MATLAB/Simulink Toolbox and Apache Spark, a powerful distributed processing framework. Finally, hyperspectral sensor datasets are processed using the MATLAB Hyperspectral Toolbox and machine learning libraries in Apache Spark to demonstrate the additional capabilities that modern computing architectures enable.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2017
- Accession Number
- AD1053363
Entities
People
- Ryan P. Mcardle
Organizations
- Naval Postgraduate School