Thwarting Location Inference Attacks in Database-Driven Spectrum Sharing

The use of geolocation databases (GDBs) for enabling spectrum sharing has been mandated by the Federal Communications Commission in the U.S. TV band and the 3.5 GHz band, and it is very likely to be adopted for spectrum sharing in other bands as well. Although using GDBs is a practical approach for enabling spectrum sharing, it raises a potentially serious operational security problem, especially when some of the primary users (PUs) are federal government entities, including military users. In this paper, we show that malicious queries can readily infer the locations of PUs by making multiple queries to the database even if the database's responses to queries do not directly reveal such information. To address this issue, we propose a perturbation-based optimal obfuscation strategy that can be implemented by the GDB to preserve the location privacy of PUs. The proposed obfuscation strategy is optimal in the sense that it maximizes PUs' location privacy while ensuring that the expected degradation in the secondary users' utility due to obfuscated responses does not exceed a threshold. Using results from extensive simulations, we demonstrate the efficacy of the proposed scheme in countering location inference attacks, including the inference strategy described in the paper.