Advanced Radar Techniques for Planetary Exploration: 3D Subsurface Imaging of Mars and Near-Earth Asteroid Characterization

Planetary radar sounding enables the investigation of subsurface structures and three-dimensional characterization of celestial bodies that are otherwise inaccessible through optical observations alone. This thesis addresses advanced radar signal processing techniques for two complementary applications: subsurface imaging of Mars through MARSIS (Mars Advanced Radar for Subsurface and Ionosphere Sounding) data, and three-dimensional shape reconstruction of near-Earth asteroids using ground-based radar observations. The theoretical foundation is established through a comprehensive treatment of electromagnetic wave propagation, surface scattering mechanisms, and radar equation fundamentals. Particular emphasis is placed on digital signal processing techniques for radar sounding, including matched filtering, pulse compression, and synthetic aperture radar (SAR) principles. The framework extends to advanced focusing algorithms, with detailed analysis of both time-domain and frequency-domain backprojection methods, inverse scattering theory, and tomographic reconstruction approaches. This theoretical synthesis is necessary given the vast and fragmented nature of radar imaging theory across different application domains. The investigation of Mars subsurface focuses on MARSIS data from the south polar region, specifically the Ultimi Scopuli area where previous studies identified anomalously bright basal reflections potentially consistent with subsurface liquid water at an apparent depth of 3 km. Through comparison with simulated data, a remnant phase term was discovered causing incoherent integration of orbital data. This led to a complete revision of the raw MARSIS processing chain, identifying noise sources including demodulation artifacts, orbital radial velocity variations, and ionospheric phase delays. The core contribution consists in the development and implementation of a three-dimensional frequency-domain backprojection algorithm, specifically optimized for MARSIS acquisition geometry and data characteristics. Applied first to a subset of 3 orbits selected for their positive focusing gain, the algorithm produces a volumetric reconstruction of the subsurface reflectivity that is qualitatively improved with respect to the uncorrected case, with spatially separated bright spots consolidating into a single, geometrically connected reflector after phase and noise correction. The reconstruction pipeline was subsequently extended to the full set of 27 noise- and phase-corrected orbits over the Ultimi Scopuli region. The resulting 3D volume reveals a coherent, elongated basal reflector at an apparent depth of 3km whose geometry and spatial extent are qualitatively consistent with the map of anomalously high basal dielectric permittivity reported by [1], providing qualitatively and independent three-dimensional geometric evidence for the spatial continuity of the candidate subsurface structure. Ground-based radar is a powerful tool for the characterization of Near-Earth Objects (NEOs), using both continuous-wave (CW) and delay-Doppler imaging techniques. Contributing to the European Space Agency pilot project "NEO Observation Concepts for Radar Systems" and to the subsequent activities, radar observations have been conducted using several radio telescopes in Europe as receivers, including the 32m "G. Grueff" Medicina antenna. For asteroid 2005 LW3, high-resolution integrated spectra revealed a binary system through the detection of a secondary spectral peak. In this work, a progressive implementation of three-dimensional shape reconstruction algorithms was developed to address scenarios where no range information is available, evolving from two-dimensional convex hull models through rotational ellipsoid fitting, radial and tangential perturbation models, culminating in a sophisticated adaptive vertex model capable of locally augmenting vertex density by optimizing residuals against spectral data. The reconstruction demonstrates that accurate 3D characterization is achievable using only Doppler data from CWobservations. Additionally, this work includes pioneering efforts in Very Long Baseline Interferometry (VLBI) for asteroid radar observations. A complete feasibility study was performed for the observation of asteroid 2025 FA22, including SNR estimates for each telescope and baseline, as well as simulations of observational visibility conditions. This represents one of the first VLBI radar observations of an asteroid, with successful detection confirmed by Lovell telescope data, establishing a foundation for future interferometric radar campaigns. This research advances planetary radar sounding through the development of robust signal- processing algorithms tailored to challenging acquisition scenarios for subsurface exploration on Mars, while it also implements innovative radar-based 3D modeling algorithms for near-Earth objects, supporting both scientific investigation and planetary defense applications.