Approved for public release; distribution in unlimited. / Synthetic aperture sonar beamforming and signal processing relies on properly steering and focusing the aperture beam pattern in order to co-phase all the received signals. Due to the effects of motion in the synthetic aperture sonar problem, the propagation path between the transmitter, discrete point scatterer, and the receiver is time varying. Traditionally, simple approximations are used to determine these propagation ranges and angles of incidence and scatter. Methods to determine these ranges and angles exactly may significantly improve array gain and, therefore, target detection. This thesis investigates improvements to SAS signal processing algorithms using exact methods for the calculation of the time-varying ranges between transmitter and discrete point scatter, and between discrete point scatter and receiver, and the phase angle of the scattered acoustic signal incident upon the receiver. Using computer simulations, exact range and angle calculations were performed for different scenarios and compared to ranges and angles determined using standard approximations. The exact ranges were then used to determine incident phase, and were again compared to the approximate methods. Comparison of the exact and approximate methods was based on range estimation error and percentage error. Improvements in synthetic aperture array gain using exact phase weights based on exact, time-varying range solutions are proposed. / http://hdl.handle.net/10945/1142 / Lieutenant, United States Navy

This dissertation proposes a learning-based, physics-aware soil moisture (SM) retrieval algorithm for NASA’s Cyclone Global Navigation Satellite System (CYGNSS) mission. The proposed methodology has been built upon the literature review, analyses, and findings from a number of published studies throughout the dissertation research. Namely, a Sig- nals of Opportunity Coherent Bistatic scattering model (SCoBi) has been first developed at MSU and then its simulator has been open-sourced. Simulated GNSS-Reflectometry (GNSS-R) analyses have been conducted by using SCoBi. Significant findings have been noted such that (1) Although the dominance of either the coherent reflections or incoher- ent scattering over land is a debate, we demonstrated that coherent reflections are stronger for flat and smooth surfaces covered by low-to-moderate vegetation canopy; (2) The influ- ence of several land geophysical parameters such as SM, vegetation water content (VWC), and surface roughness on the bistatic reflectivity was quantified, the dynamic ranges of reflectivity changes due to SM and VWC are much higher than the changes due to the surface roughness. Such findings of these analyses, combined with a comprehensive lit- erature survey, have led to the present inversion algorithm: Physics- and learning-based retrieval of soil moisture information from space-borne GNSS-R measurements that are taken by NASA’s CYGNSS mission. The study is the first work that proposes a machine learning-based, non-parametric, and non-linear regression algorithm for CYGNSS-based soil moisture estimation. The results over point-scale soil moisture observations demon- strate promising performance for applicability to large scales. Potential future work will be extension of the methodology to global scales by training the model with larger and diverse data sets.

A study is undertaken of the numerical reconstruction of acoustic and electromagnetic (EM) localized waves (LWs). The latter are carrier-free ultra-wideband pulses characterized by large focusing depths and extended ranges of localization. Special emphasis is placed on finite energy LWs that can be generated by dynamic aperture antennas with independently addressable elements. The reconstruction techniques are based on Huygens and Rayleigh-Sommerfeld integral I and II representations, both in the time and frequency domains. In contradistinction to the Weyl representation,they lend themselves to the physical realization of space-time aperture sources capable of generating localized wave solutions propagating away from the aperture plane. A detailed comparison of the three reconstruction techniques has been carried out in connection with LW solutions to the scalar wave equation, especially with respect to their handling of acausal components incorporated in the aperture excitation fields. In addition, a study is presented of the characteristic properties of LWs propagating through dispersive media modeled by the Klein-Gordon equation. It is demonstrated that contrary to expectation, the depletion of the spectral components of the LW Klein-Gordon field may be slower than that associated with the free space scalar field. Previous work by Power et al. [73] is extended by studying the acoustic bistatic scattering of a modified power spectrum (MPS) pulse from rigid and compressible spheres. The analysis allows the extraction of the radius of a sphere from the backscattered data. Finally, a special class of electromagnetic (EM) LWs, referred to as azimuthally polarized X waves (APXWs), is derived and their reconstruction is addressed, both in the time and frequency domains. / Ph. D.

L’étude de la diffusion des ondes électromagnétiques par une surface rugueuse aléatoire est de première importance dans de nombreuses disciplines et conduit à diverses applications notamment pour le traitement des surfaces par télédétection. En connaissant les modes de rétrodiffusion, on peut détecter la présence de la rugosité aléatoire indésirable de la surface de réflection telle que le réflecteur d'antenne et par conséquent trouver un moyen de corriger ou compenser les erreurs de phase. Cette thèse porte sur l’obtention de l'humidité du sol de surface à partir de mesures radar. La description de la surface rugueuse de façon aléatoire est présentée, suivie par les interactions d'ondes électromagnétiques avec les média. En particulier, un modèle d'équation intégrale avancé (AIEM) est introduit. La validité du modèle AIEM, qui est adopté comme modèle de travail, se fait par une large comparaison avec des simulations numériques et des données expérimentales. On analyse également les caractéristiques des configurations radar bistatique et on étudie la sensibilité de la diffusion bistatique à l'humidité du sol et à la rugosité de surface et, dans le même temps, le cadre de la détermination de l'humidité du sol à partir de mesures radar utilisant un réseau de neurones à base de filtres de Kalman récurrents est présenté. La formation du réseau et l'inversion des données sont décrits. / Electromagnetic waves scattering from a randomly rough surface is of palpable importance in many fields of disciplines and bears itself in various applications spanned from surface treatment to remote sensing of terrain and sea. By knowing the backscattering patterns, one may detect the presence of the undesired random roughness of the reflection surface such as antenna reflector and accordingly devise a means to correct or compensate the phase errors. Therefore, it has been both theoretically and practically necessary to study the electromagnetic wave scattering from the random surfaces. This dissertation focuses on the retrieval of surface soil moisture from radar measurements. The description of the randomly rough surface is presented, followed by the electromagnetic wave interactions with the media. In particular, an advanced integral equation model (AIEM) is introduced. The validity of the AIEM model, which is adopted as a working model, is made by extensive comparison with numerical simulations and experimental data. Also analyzes the characteristics of the bistatic radar configurations and dissects the sensitivity of bistatic scattering to soil moisture and surface roughness of soil surfaces. Meanwhile presents a framework of soil moisture retrieval from radar measurements using a recurrent Kalman filter-based neural network. The network training and data inversion are described in detail.

Dans le cadre de la mission spatiale SMOS (Soil Moisture and Ocean Salinity), nous présentons dans cette thèse une nouvelle approche numérique de modélisation du calcul de l’émissivité et du coefficient bi-statique de systèmes forestiers sol-litière en Bande L. Le système sol-litière est représenté par deux couches diélectriques 3D comportant des interfaces rugueuses, une démarche qui n’apparait pas actuellement dans la littérature. Nous validons notre approche pour une seule couche en comparant les simulations de l'émissivité avec celles produites par la méthode des moments et des données expérimentales. A partir de ce nouveau modèle, nous évaluons la sensibilité de l’émissivité du système sol-litière en fonction de l’humidité et de la rugosité de la litière. Ce nouveau modèle permettra de créer une base de données synthétiques d’émissivités calculées en fonction de nombreux paramètres qui contribuera à améliorer la prise en compte de la litière dans l'algorithme d’inversion des données de la mission spatiale SMOS. / In the context of the SMOS (Soil Moisture and Ocean Salinity) mission, we present a new numerical modelling approach for calculating the emissivity and bistatic scattering coefficient of the soil-litter system found in forests, at L-band. The soil-litter system is modelled as two 3-dimensional dielectric layers, each with a randomly rough surface, which to our knowledge has not previously been achieved. We investigate the validity of the approach for a single layer by comparing emissivity simulations with results of Method of Moments simulations, and experimental data. We then use the approach to evaluate the sensitivity of the soil-litter system as a function of moisture content and the roughness of the litter layer. The numerical modelling approach which has been developed will allow us in the future to create a synthetic database of the emissivity of the soil-litter system as a function of numerous parameters, which will contribute to validating and improving the inversion algorithm used by the SMOS mission to retrieve soil moisture over forests.