Profiling tropospheric refractivity in real time, based on a relevance vector machine and single ground-based GPS receiver

Abstract

A method is developed for retrieving atmospheric refractivity profiles in real time from the slant path tropospheric delay (STD) at low elevations (<5°) of a single ground-based Global Positioning System (GPS) receiver. The technique for retrieving atmospheric profiles proposed herein is based on a relevance vector machine (RVM), which is a sparse approximate Bayesian kernel method. The RVM network is trained by using historical radiosonde data, while the simulated STDs are calculated using the ray-tracing technique. The inputs of the network are surface meteorologic parameters and differences of the STDs at various elevations, while the outputs are tropospheric refractivity profiles for specific layers. The feasibility of this method is verified by both simulation and experiment. The STDs of the GPS receiver are obtained in the experiment by the undifferenced precise point position (PPP) method with International GPS Service (IGS) ultra-rapid products, which means that this method may be applied in real time. This article proposes a new method for retrieving atmospheric refractive profiles in real time at low cost and with high efficiency.