Permittivity Estimation in Ray-Tracing Using Path Loss Data Based on GAMP

Abstract

With the electromagnetic (EM) parameter demands of wireless link prediction such as ray-tracing (RT), using field wireless measurements to calibrate environmental EM parameters becomes a promising solution to obtain accurate EM parameters in a given area. In this letter, we propose a modified generalized approximate message passing (GAMP) algorithm with trust region to estimate permittivity parameters in RT model based on path loss data. The aim of trust region is to constrain the parameters and reduce the inaccuracy while dealing with messages updated in GAMP for nonlinear estimation. Numerical results demonstrate that the proposed algorithm outperforms GAMP without trust region in two aspects: the accuracy of permittivity estimation, as well as the errors in path loss prediction utilizing the calibrated parameters.