Pre-deployment performance assessment of device-free radio localization systems

Device-free localization (DFL) systems have emerged in the last years as a powerful technology for tracking mobile targets without the need of radio tags. Perturbations induced by moving objects on the electromagnetic (EM) wavefield generated by a dense wireless network are measured and processed by the DFL system to track target trajectories. Despite several solutions have been explored in the literature, mainly based on fingerprinting approaches, a deep understanding of body-induced effects on the EM fields for target tracking is still missing as well as reliable predictive models for pre-deployment accuracy assessment of DFL systems. The paper makes a first attempt towards the definition and validation of a novel predictive tool that is general enough to be applied to DFL systems with any kind of RF interface, network topology and connectivity degree. An analytical diffraction model is exploited to predict the effect of a human body on the received signal strength field over all the available links and compute fundamental limits to the DFL positioning accuracy. The proposed tool is tailored for 2D human body localization and validated by experimental trials in an indoor environment.