PCA-Based inversion of WiFi signal for robust device-free indoor target detetion

The spatial and frequency diversity introduced by the recent WiFi standards based on orthogonal frequency-division multiplexing is exploited for the robust wireless detection of device-free targets in indoor environments. The arising inverse problem is solved by decomposing the channel state information by means of a customized strategy based on the principal component analysis (PCA) for extracting target-dependent, components from the WiFi signal. The experimental validation of the proposed solution pointed out a robust target detection with failure rates lower than 3 [%] using a single wireless link already deployed in a real office test, field.