News

GNSS tomography resolution matrix spread - another overly complicated solution to simple problem?

05-05-2022

GNSS tomography is a technique that is looking into using observation penetrating the medium from multiple of angles to establish the 3D structure of that medium. This is inverse Radon transform, used in many - mostly medical applications (PET/CT scanning). Similar is done with GNSS observations that penetrates the atmosphere from multiple angles, accumulated impact of the atmosphere on the signal phase is recorded at the antenna point, and then inverted to produce 3D structure of atmosphere. Fairly simple, however the inversion is not always successful and the resultant 3D model becomes blurry or contains a lot of outliers.

When it happens? Why it happens? How to test for potentially unstable solution before running Radon inversion?

The team of UPWr (Witold Rohm) and TU Vienna researchers (Zohreh Adavi and Robert Weber), just published a paper in Journal of Geodesy that is solving a problem of tomography instable results – without the need of finding laborious inversion. The paper “Pre-analysis of GNSS tomography solution using the concept of spread of model resolution matrix” is providing effective methodology to test the tomography effectiveness in based on the spread of the resolution matrix. This method will effectively work in real-time solutions to accumulate enough observations epochs for stable solution, in testing the impact of adding new observations (e.g. radio occultation observations), preventing introducing spurious or non-effective observations.
More details are available here - Journal of Geodesy