Application of RF-DNA Fingerprinting Techniques to ICOM Radio Satellite Communication
The full-text download link for this item is available to Department of Defense personnel only at this time.
Abstract
This research addresses if RF-DNA techniques can be utilized to correctly identify and classify the signals coming from six different communication ground-stations. It performs this analysis through the fingerprinting of features in the Time Domain as well as in the Spectral Domain. The classification is performed utilizing MDA/ML and Random Forest Classifiers. Additionally, it introduces a Time/Spectral Domain combined fingerprint set and discusses its potential applicability. Included in the scope of the research is an analysis of the performance of the RF-DNA classification with regards to the amount of oversampling performed on the signal collection. Time/Spectral Domain fingerprints yielded a higher classification performance than the two baseline fingerprinting method. Higher classification performance can be gained through oversampling, however there is a trade-off when considering calculation time and memory. Overall, RF-DNA techniques have high applicability to satellite communications and could complement existing security methods.