Spoofing Attacks against FM Indoor Localization

Abstract

FM-based indoor localization identifies the location of a user by looking at the received signal strength (RSS) at the user's location. We find an effective spoofing attack that can remotely alter RSS of a target user, leading to the incorrect estimation of the user's location. By analyzing the features of the received signals and FM signals in the frequency domain, we propose a defense method to deal with this attack. The proposed method contains two levels of detections. Specifically, the first level of detection distinguishes between the normal signal and the attack/noise signal, and the second level of detection finally detects the existence of the attack signal. We perform real-world experiments on Universal Software Radio Peripherals (USRPs) to spoof a target location to four different locations. The experiment results show a promising performance of the proposed defense method with a false negative rate of 3.9% and a false alarm rate of 6.4%, respectively.