Testing of ground penetrating radar at the lunar analog facility (LUNA): A robust reference frame for planetary subsurface exploration

Introduction

In the last decade planetary science has seen an emergence of in-situ, i.e. rover-based, radar subsurface investigation missions on planetary bodies, such as Chang'E 3 and 4 on the Moon as well as the Radar Imager for Mars’ Subsurface Experiment (RIMFAX). Commonly, the geologic stratification and material composition shall be determined, including special targets such as water ice or cavities. 

In the meantime, the German Aerospace Center (DLR) and ESA have provided favorable testing conditions through the lunar analogue facility (LUNA), a site that allows human and robotic training in a Moon-like setting [Casini 2020]. The LUNA site features a large-scale pit with a deep (3.0 m) and shallow section (0.6 m), filled with the well characterized EAC-1A lunar regolith simulant. The pit naturally represents a test laboratory for shallow geophysical exploration techniques such as GPR and their inclusion in in-situ planetary exploration campaigns [Knapmeyer 2025].

We employ techniques commonly used on planetary in-situ radar missions. However, without ground-truth information in space it is difficult to disambiguate findings, in particular high-resolution observations such as small target detection and local material parameters estimation. On the contrary, LUNA is both well characterized and proficiently Moon-like, allowing us to validate these methods in conjunction with new instruments.

Figure 1: Left: LUNA hall at arrival. Middle: Buried target sites before regolith infill. Operators need to wear protective gear in the LUNA hall. Right: Example of buried boulder target

Figure 2: Left: Suspended antenna in a low height mode over buried targets. Right: Rover-based radar deployment

Measurements

The campaign of November 2024 focused on the shallow area of 0.6 m depth since the deep area was still subject to filling. Various structures were emplaced into the regolith that model possible targets inspired by Mars and Moon studies [e.g. Casademont 2023]. The targets are comprised of three buried boulders in the depths deep, shallow and surface, a micro impact crater, a duricrust patch, a buried cavity, air-targets, that is targets above ground, both in-drive and out-of-plane as well as a metal plate at maximum depth for reference. Fig. 1 and 2 show the measurement context.

The radar electronics are a Technical University of Dresden built IQ LFMCW 0.3 – 6 GHz full-polarimetric radar. Different antennas have been employed in different heights, with and without targets, with targets in nadir and out-of-plane direction. A precursor campaign sounded the LUNA hall without regolith infill in a similar manner. The set of antennas include a crane-suspended dipol antenna as in Fig 1. and theWater Ice Subsurface Deposit Observation on Mars (WISDOM) antenna deployed directly on the ground [Benedix 2024]. Horizontal and Vertical polarization have been switched for forward and backward traverse along the same measurement path.

Together with the radar elements a set of secondary sensors was deployed. It contained an ultrasound positioning system (MarvelMind), a stereo depth imaging camera pointed in nadir direction for the suspended antenna setup (Orbbec Astra) and a Lidar (R2000) in the x-y plane. Depth camera and Lidar are employed to characterize the site with a high-resolution digital elevation model for subsequent surface clutter estimation. Apart from measurement geometry, the important ground-truth are laboratory dielectric parameters of the EAC-1A substrate. They were previously characterized at room temperature and uncompacted (rho=1.72 g/cm3) for frequencies above 400 MHz by [Ramos Somolinos 2022]. The authors report ε′ = −0.0432f + 4.0397 and tan δe = −0.0015f + 0.0659 (linear fitting). While they describe non-magnetic behavior of EAC-1A, a handheld magnet attracted significant amount of EAC-1A during this campaign. Nonetheless, their characterization serves as temporary ground-truth material parameters until further research is conducted. Volumetric water content sensors installed shortly after the campaign show values 2-5% for the top 4 cm of regolith. 

First results

Standard minimal processing includes data curation (sounding localization, repetitive measurement removal, homogenization of sensor clocks, healing of bin-shifts in recording), a windowed Fourier Transform along the soundings and a moving average subtraction to remove the constant background component. Given the rich experimental setup, we focus here on a first assessment regarding target detection.

Fig 3. shows data of the profile where the antenna was suspended with 0.45 m ground clearance at profile start. In the upper subplot, hyperbolic structures can be observed, that are typically associated with localized targets of small extent compared to the wavelength. They mostly coincide with known target positions, for instance the buried reflector target at 2.5m, 28 ns as well as the surfacing targets and air-targets from the micro crater onwards to the right. Hyperbolic signatures are also present around the buried boulder positions, yet their identification is not without ambiguity in standard processing. Surface effects introduce a high degree of hyperbolic scattering

The phase image shows a horizontally undulating band of lower phase variation, which coincides in traveltime with the concrete floor of the LUNA area. Note that the concrete layer flatness will be overprinted with the undulating depth of the regolith infill.

Figure 3: Standard processing, feedpoint time around 18 ns. Top: real part of signal with short term background removal. Bottom: IQ-phase of signal with global BGR

Discussion & Outlook

The experiments show promising results with respect to detectability of typical planetary targets. Tehy also shows that detection of buried targets, especially with respect to hyperbolic permittivity derivation or specific target identification, cannot be considered a trivial task, even under controlled circumstances. Radar instruments need to be fine-tuned and possibly secondary sensors integrated for surface clutter characterization.

References

Casini et al., 2020, Lunar analogue facilities development at EAC: the LUNA project, 10.1016/j.jsse.2020.05.002

Knapmeyer et al., 2025, First campaigns and future developments in the LUNA Moon analog facility, 10.5194/egusphere-egu25-19008

Casademont et al., 2023, RIMFAX Ground Penetrating Radar Reveals Dielectric Permittivity and Rock Density of Shallow Martian Subsurface, 10.1029/2022JE007598

Benedix et al., 2024, The ExoMars 2028 WISDOM antenna assembly: Description and characterization, 10.1016/j.pss.2024.105995

Ramos Somolinos et al., 2024, Electromagnetic Characterization of EAC-1A and JSC-2A Lunar Regolith Simulants