WISDOM Calibration Methodology and Instrument Transfer Function Impact on Surface Permittivity Measurements and Depth Resolution
Résumé
WISDOM, the “Water Ice Subsurface Deposition on Mars” Experiment, is as Ground Penetrating Radar (GPR) part of the 2020 ESA-Roscosmos ExoMars Rover payload. It will be the first fully polarimetric planetary GPR tasked at probing the subsurface of Mars.
The WISDOM GPR will give insight in the geological structure of the shallow subsurface at an unprecedented (centimetric) resolution with up to 3m of penetration. WISDOM will support the drill operations by identifying locations of high scientific yield and low risk.
When reliable quantitative measurements are the expected instrument outcome, the key challenge is calibration. Regarding GPR data this is especially true when permittivity values (e.g. hinting at subsurface composition) are to be derived from raw measured data. Assuming a model where the measured radar data is a convolution of the instrument transfer function and the channel impulse response from surface and subsurface, knowledge of the instrument transfer function has a significant influence on the calibration and furthermore, radar depth resolution and effective instrument dynamic range.
In order to estimate the instrument transfer function and establish the instrument calibration, multiple reference measurements are performed using defined metallic reflectors in multiple known distances. To compensate, or eliminate, disturbing propagation paths from the measurement environment, the radar data is sorted to maximize the low-rank property of the measurement data matrix, transformed in an affine domain using a singular value decomposition and filtered in the affine domain.
The calibration data processed in this manner is used as reference in the permittivity estimation, while the determined instrument transfer function is used for the extraction of the channel impulse response from the measured radar data.
We demonstrate the feasibility and advantages of this calibration and estimation approach on simulated data, on data obtained from measurements under laboratory conditions, as well as data from field tests.