Differentially Private Target Localization Based on DOA Measurements

Target localization in open environments faces challenges in preserving the privacy of target, e.g., the location of user. In this paper, we address this challenge by proposing a target localization framework that integrates differential privacy and the improved localization algorithm, which is carried out based on the direction of arrival (DOA) measurements. Our approach begins with an input perturbation mechanism, where the elaborate noise is injected into measurements to enforce strong privacy guarantees. To counteract the performance degradation caused by parameter mismatch, an improved target localization algorithm is proposed exploiting the statistical properties of privacy preserving measurements. Additionally, the privacy budget allocating (PBA) approach is also proposed in this paper, which is derived by minimizing the Cramér-Rao lower bound (CRLB) of target localization under the region-specific privacy constraints. Leveraging this foundation, the problem of joint PBA and sensor placement is further considered in this paper, and such a problem is solved via an alternating optimization approach. The proposed algorithms in this paper can balance the differential privacy of transmitted data as well as the accuracy of target localization. The performance of proposed algorithms is verified via simulation experiments.