A Lightweight Detection of the RFID Unauthorized Reading Using RF Scanners

Many RFID tags store valuable information that can easily be subject to unauthorized reading, leading to system security and privacy risks. The detection methods existed are not only complex and impractical, but also unable to extract more information about the abnormal signal. In this paper, we propose a lightweight detection approach for the unauthorized reading without affecting the operation of RFID systems. Such an approach contains three parts: RF signal scanner, signalevent model construction and abnormal feature extraction. In particular, we design and implement a RF scanner to acquire RF signals and measure RSSI values. After that, we build a signal-event model to analyze how the RSSI value is related to the RFID event. The detection of unauthorized reading is to investigate the deviation of observed RSSI values from their expected values. Finally, we extract and separate abnormal RSSI values to estimate the risk of unauthorized reading. The primary experimental results show that our approach can achieve high prediction accuracy in detecting unauthorized reading and make better performance in extracting abnormal features.