Frequency-Based Inversion of a Single Wireless Link for Indoor Passive Target Detection

In this work, the wireless detection of transceiver-free targets is formulated as an inverse problem and solved using the received signal strength of existing wireless devices. The main goal of the proposed solution is to opportunistically exploit a single wireless link established between two commercial wireless access points for the real-time detection of humans entering the monitored domain. The proposed computational method exploits both an empirical mode decomposition for denoising and a discrete Fourier transform for filtering the received signal strength. A reduced set of highly informative features are used for the automatic learning of a support vector machine classifier and the prediction of target absence/presence probability. A preliminary experiment carried out in a real indoor test site is presented to assess the advantages and limitations of the proposed inversion method.