Date of Award
Fall 2023
Embargo Period
5-25-2028
Access Type
Thesis - ERAU Login Required
Degree Name
Master of Science in Engineering Physics
Department
Physical Sciences
Committee Chair
Kshitija Deshpande
First Committee Member
Karl Martin
Second Committee Member
Thomas Yang
College Dean
Peter Hoffman
Abstract
Satellite communications can be interfered from intentional or unintentional sources, which causes the degradation of their communication links. This interference further leads to a reduction of data passed through the communication link and reduction of access times. Due to these risks, the interference events need to be detected in an autonomous fashion to reduce satellite operator error and workload. After detection has occurred, actions can be taken to mitigate the effects of the interference. Additionally, the detection method needs to be detected on-orbit to eliminate the need for extra data to be sent through the communication link. Because of the systems engineering limitations in the build and launch of a satellite, this leads to size, weight and power requirements for the detection method similar to a passive sensing method that cannot take up the resources such as that of a dedicated payload. This thesis focuses on the detection aspect and neglects the mitigation aspect of the SATCOM interference problem. The methods section details the design of a software defined radio test bed for the creation of training data, as well as the processing of this data to be made ready for machine learning models. Two edge-based passive sensing methods that are conventional spectrum sensing cognitive radio techniques and a machine learning approach while comparing the results of several classifier models. These detection methods are evaluated using a binary classification metric of interfered and not interfered, with the run times presented and accuracies consistently above 0.99. The features used for these detection approaches is the telemetry output from a command receiver, enabling the method to be run at a significantly lower rate without aliasing and without the loss of information.
Scholarly Commons Citation
Brock, Arthur, "Machine Learning and Spectrum Sensing Cognitive Radio for Satcom Interference Detection" (2023). Doctoral Dissertations and Master's Theses. 786.
https://commons.erau.edu/edt/786