Principal Investigator: Edward Delp
Given the wide use of RF devices for applications ranging from data networks to wireless sensors, it is of interest to identify the types of devices that are located in an environment. In order to locate and characterize RF devices, the environment must be probed. This becomes the problem of determining the properties of an RF circuit by sending it a carefully designed signal and examining the returned signal. The device is then authenticated by identifying certain characteristics of the return signal. Identification through “forensic characterization" means identifying the type of RF device, make, model, configuration, and other characteristics of the device based on observations of the data that the device produces.
The characterization framework is used to classify signals emitted from an RF device that are generated in response to a probe signal. Once transmitted into the environment, the probe signal is received by the antenna of the RF device and sent to several circuit components such as filters and amplifiers. Each circuit component modifies the probe signal, thereby embedding information into the signal. The embedded information is unique to each circuit component and when input into a nonlinear component such as an amplifier, a percentage of the probe signal is reflected. The reflected signal is re-radiated from the RF device and measured. Once measured, a forensic signature is extracted from the re-radiated signal and characterized by a classification system.
Other PIs: Kevin G. Gard (North Carolina State University) Andreas Cangellaris (University of Illinois at Urbana-Champaign) Lawrence Carin (Duke University) William J. Chappell (Purdue University) Michael Steer (North Carolina State University)
A. K. Mikkilineni, D. King-Smith, S.B. Gelfand, and E.J. Delp, "Forensic Characterization of RF devices," First IEEE International Workshop on Information Forensics and Security, WIFS 2009, pp.26-30, 6-9 Dec. 2009.
Keywords: forensic characterization, RF device