Machine Learning-Based Network Communication Protocol Optimization for the Internet of Things
As smart devices continue to see perpetually growing use, immense volumes of data are collected, stored and processed. The need for an efficient and secure communication system has grown more crucial than ever. The student will investigate how to utilize key features and characteristics of Communication protocols for designing machine learning algorithms is used to iteratively generate optimized communications networks for connected devices such as the Internet of Things (IOT). Fitness of the machine learning-generated communication protocol will be based on the protocol's speed of initiating and terminating connections, as well as the average error rates of communication. data scrambling, originally for minimization of electromagnetic interference, will be reverse engineered and re-purposed for security and encryption, whose overhead in data transmission will be optimized via supervised machine learning.
- Designing and developing hardware implementations - Understanding discrepancies between the theoretical and practical via technological limitations - Learning to use industry-standard equipment (TLA715 6GHz digital logic analyzer,TDS5104B digital phosphor oscilloscope, Virtex-5 ML510 FPGA development board) - Augmenting programming knowledge and skills (Python, C, Java, Matlab, Shell, Assembly)
Course level knowledge in ELE404, COE428, MTH514, COE538, ELE604, and ELE635 are requirement. Strong interest in learning about digital systems, digital signal processes and data analysis. This project requires the basic knowledge of Internet of Things, machine learning, communication protocols, and mobile application development. Additionally, the student requires a basic understanding of optimization problems, a strong foundation in basic communications theory, and an in-depth review of analysis tools,algorithms and programming.
Reza Sedaghat : Machine Learning-Based Network Communication Protocol Optimization for the Internet of Things | Monday April 1st 2019 12:43 PM