Performance Evaluation of Carrier-Frequency Offset as a Radiometric Fingerprint in Time-Varying Channels.

Area of Science:

• Wireless communication security
• Physical layer authentication
• Signal processing

Background:

• Physical layer attributes are increasingly used for wireless device authentication.
• Carrier frequency offset (CFO) shows promise for device identification in static environments.
• The performance of CFO in dynamic, high-mobility scenarios remains largely unexplored.

Purpose of the Study:

• To experimentally evaluate the accuracy of CFO for wireless device identification in time-varying channels.
• To assess the impact of vehicular mobility on CFO-based authentication effectiveness.

Main Methods:

• Deployment of a software-defined radio (SDR) testbed for real-world data collection.
• Collection of CFO values in vehicular setups under varying mobility conditions.
• Training of machine learning (ML) classifiers using collected CFO data for device identification.

Main Results:

• CFO achieved high accuracy (97%) for device identification in low-mobility scenarios.
• Accuracy degraded to 72% at higher mobility (35 miles/h).
• Time-varying channel conditions in high-mobility scenarios negatively impact pilot signal quality for CFO detection.

Conclusions:

• CFO is a viable authentication feature for low-mobility wireless devices.
• High mobility significantly reduces the effectiveness of CFO for distinguishing transmitters.
• The study highlights limitations of CFO-based authentication in dynamic environments.