Synthetic Generation of Radio Maps for Device-Free Passive Localization

In this paper, we present the design, implementation, and evaluation of a system that automatically constructs accurate radio maps for device-free WLAN localization systems. The system is capable of generating deterministic and probabilistic radio maps for localization systems. Our system uses 3D ray tracing enhanced with the uniform theory of diffraction (UTD) to model the electric field behavior and the human shadowing effect. We present our system architecture and describe the details of its different components. We also propose an optional module, location-0 correction, that can significantly enhances the system accuracy and reduces its dependence on the 3D model details by using just one signal strength sample. Our experiments in a real testbed show that the predicted signal strength differs from the measurements by a maximum average absolute error of $2.77$ dB achieving a maximum localization error of $3.13$m and $2.84$m for both the deterministic and probabilistic radio maps, respectively. In addition, the results show that our system is not sensitive to the 3D model details.