A Joint Rogue Access Point Localization and Outlier Detection Scheme Leveraging Sparse Recovery Technique

With the pervasive deployments of Access Point (AP) in Wireless Local Area Network (WLAN), rogue AP has emerged as such a large threat to user's privacy, that it is expected to be detected and located accurately. Hence, in this paper, we propose a novel rogue AP localization scheme leveraging sparse recovery technique, which consists of three steps: 1) Coarse localization, which is the result of comparing the online records with clustered fingerprint database. A novel Object Weighting Affinity Propagation (OWAP) clustering method is proposed to group the offline fingerprints. When computing the similarities, unlike traditional affinity propagation clustering method which views each object equally, we utilize the prior physical coordinates information to assign weight to each object. 2) Compressive sensing (CS) kernel optimization, in which the minimum number requirement of monitors in localization system is deduced through analyzing the problem formulation theoretically, and an Equiangular Tight Frame (ETF) based monitors selection scheme is presented to achieve higher location accuracy. 3) Joint fine rogue AP localization and outlier detection through a formulation of an improved CS based sparse recovery model. It could localize the rogue AP and identify the monitors whose readings are either not available or erroneous simultaneously, which increases the localization robustness. We operate simulations as well as experiments to verify the superiority of the proposed scheme both theoretically and practically.