Comparative Analysis of RF Emission Based Fingerprinting Techniques for ZigBee Device Classification
Abstract
LR-WPAN are increasingly being fielded to complete tasks in autonomous sensor networks, industrial control systems, and other critical infrastructure. ZigBee is a versatile LR-WPAN platform that also open to risks of sophisticated bit-level attacks. PHY based security measures have been shown in previous research efforts as effective supplemental security measures that a not susceptible to bit-level attacks. This research effort intends to quantify the differences in various RF fingerprinting techniques via comparative analysis of MDA/ML classification results. The findings herein demonstrate a methodology for the generation of CB-DNA, RF-DNA, and COR-DNA fingerprints. The results show that CB-DNA generated fingerprints had the highest mean correct classification rates followed by COR-DNA and then RF-DNA in most test cases and especially in low Eb/N0 ranges, where ZigBee is designed to operate.