A Min-Max Optimization-Based Approach for Secure Localization in Wireless Networks

Range-based localization of a target device in wireless networks in the presence of malicious attackers that tend to disrupt the localization process by counterfeiting (spoofing) their radio measurements is addressed in this work. In contrast to state-of-the-art methods, that assume that all devices participating in the process are non-malicious in the beginning, we here tackle the problem from the opposite perspective. All devices are treated as malicious at first, and, by assuming that an upper-bound on the attack intensity is (imperfectly) known a priori, the worst-case scenario is studied, from which two novel estimators are derived. The first approach is based on convex relaxation and leads to a robust second-order cone programming (R-SOCP), while the other one assumes problem reformulation as a robust generalized trust region sub-problem (R-GTRS). Received signal strength (RSS) scenario is in the main focus, but an adaptation of the new approach to a general range-based setting is presented as well. The proposed min-max approach is validated though computer simulations, where it showed its worthiness by outperforming the state-of-the-art approaches and offering a more reliable (secure) solution to the problem. Finally, it is worth mentioning that a theoretical analysis on the detection performance is also included in the work.