Investigations into the role of entropy-selected RF-DNA fingerprint features on ID-verification performance in the presence of rogue emitters

The Internet of Things (IoT) is projected to reach 30.9 billion devices by 2025. However, most lack adequate security measures against sophisticated threats. Specific Emitter Identification (SEI) is a crucial security approach for authenticating wireless emitters. This work integrates RF-DNA fingerprinting, a specialized form of SEI, with Deep Learning (DL) techniques to authenticate the identity of authorized emitters. This authentication becomes crucial in the presence of “rogue” emitters who deliberately impersonate authorized emitters using falsified digital credentials. The RF-DNA fingerprints are extracted from the entropy-selected regions within the TF representations of an emitter’s signals. The obtained results demonstrate the success of a Convolutional Neural Network (CNN) in verifying the identities of all authorized emitters at an accuracy rate of 95% or higher. Additionally, the CNN effectively detects and rejects all twelve rogue attacks with an accuracy rate of 89% or better, at an SNR of 9 dB.

Grantor dc:publisher

University of Tennessee at Chattanooga

Year dc:date.available

2025

Author dc:creator

• Mohammed, Awab

Contributors dc:contributor

• Reising, Donald R.
• Loveless, Thomas D.; Fadul, Mohamed K. M.
• College of Engineering and Computer Science

• Biometric identification--Computer networks--Security measures
• Deep learning (Machine learning)
• Entropy (Information theory)
• Pattern recognition systems
• Radio frequency identification systems--Access control

Licence

• http://rightsstatements.org/vocab/InC/1.0/

Language dc:language

English, eng

Repository record dc:identifier

https://scholar.utc.edu/theses/861

OAI identifier oai:identifier

oai:scholar.utc.edu:theses-2041