Device classification performance modeling using UWB stimulated “RFDNA” fingerprinting

Radio Frequency Distinct Native Attribute (RF-DNA) fingerprint processing is a method to extract features received from a Radio Frequency (RF) signal for device identification or verification. The concept is based on the fact that every device emits signals that have unique characteristics that can be used to distinguish specific devices from other, similar devices. The term “fingerprinting” is used because the concept is similar to human fingerprinting, where a person´s fingerprints can be used to distinguish them from another person´s. RF-DNA fingerprinting has been in development at the Air Force Institute of Technology (AFIT) since 2006 successfully demonstrating passive feature extraction from various device RF emissions such as including IEEE 802.11 WiFi, IEEE 80211.15 BlueTooth, IEEE 802.16 WiMAX, Global Systems for Mobile Communications (GSM) cell phones, and Radio Frequency Identification (RFID) emitters. To date, all of these methods have been performed in a passive manner, whereby the signals analyzed by the RF-DNA fingerprint processing algorithms were already generated by the device under normal operation. This works well for devices that normally broadcast some RF emission that can be exploited. However, a whole class of devices exist that do not normally emit RF radiation, such as properly shielded RF amplifiers, mixers, oscillators and filters that make up a typical receiver front-end. The potential exists for these devices to be classified using an “active” interrogation method, such as a radar transmission, in which the reflected energy by can be exploited by RF-DNA fingerprint analysis.