LAMP: Lightweight and accurate malicious access points localization via channel phase information

Abstract

Malicious access points (APs) have emerged as a serious security problem. In this paper, we propose LAMP, a lightweight and accurate malicious access points localization via channel phase information. We utilize a commercial wireless card for malicious APs localization with little human effort. The localization framework consists of three components: feature extraction, direction determination and localization. We first extract the available channel phase information by a linear transformation and purify eigenvalues by denoising. Then, the direction of an AP can be accurately detected by our developed self-adaptive MUSIC algorithm. Finally, the localization of the AP is realized by a triangulation algorithm with only one device placed at multiple locations. Experimental results show the median detection error of angle is about 5∘ and median localization error is about 35 cm.