الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 106 |  |  |
| | | from | 106 | | |
Abstract
Signal acquisition is a central concern in signal processing. The well-known Shannon-Nyquist theorem forms an integral part of any traditional analog to digital converter, stating that any signal must be sampled at a constant frequency that have to be at least twice the highest frequency contained in the signal in order to completely recover the signal. Compressive sensing (CS) is a relatively new paradigm within which data acquisition and data processing are combined. CS allows data to be compressed when sampled by capitalizing on the sparsity existing in many common signals. By doing this, it offers an efficient way to minimize the frequency of measurements required for the perfect signal recovery. CS has attracted a lot of attention in recent years, with thousands of scientific papers and applications emerging in fields such as direction of arrival estimation, image processing, computational electromagnetics and many more. This thesis investigates the application of compressive sensing to the synthesis of sparse and partially adaptive antenna arrays with the goal to reduce the hardware complexity and/or achieve a better performance on the radiation pattern. In the first part of the thesis, an algorithm based on CS is proposed for forming broad nulls in linear and planar adaptive phased array antenna (PAA) by adjusting the complex excitation weights of only a small number of the array elements. The algorithm jointly finds the minimum number of elements needed to be perturbed, their locations, and the necessary weight perturbations required to form the nulls. A constraint is applied to the optimization problem to ensure that the perturbation of the selected antenna elements does not cause a pointing error in the main beam. In addition, a further constraint is added to set a predefined upper limit for the array response in the sidelobe region. Simulations for linear and planar arrays have been performed. The results show that the proposed algorithm can form the necessary nulls by modifying the complex weights of only a small number of the array elements. Next, the synthesis of thinned non-uniformly excited concentric ring arrays (CRAs) using a CSbased technique is considered. Thinned CRA is achieved by switching off some elements form a fully populated CRA. Firstly, the problem is formulated as a sparse recovery problem and then relaxed to obtain a convex optimization problem that can be solved efficiently. The proposed CS-based method is compared to global evolutionary optimization-based techniques from the literature such as backtracking search optimization algorithm and firefly algorithm. The results proved the effectiveness of the proposed method compared to the previously published results from the literature. The third part of the thesis considers the synthesis of sparse and pattern reconfigurable CRAs. For sparse CRA, the array is optimized to use a minimum number of non-uniformly spaced concentric rings with uniformly spaced radiating elements in each ring. Pattern reconfigurable CRA can radiate multiple focused and/or sector beams by adjusting only the excitations of the different rings. In this thesis, the multiple patterns were optimized jointly to find the common ring radii and the different set of excitations for each pattern. This is important to ensure that all patterns share the same array structure, because optimizing the array for each pattern could result in different sets of ring radii for different patterns. Consequently, the number of rings and thus the elements required to radiate multiple patterns would increase. In the fourth part of the thesis, a method is proposed for synthesizing sparse CRAs that use a minimum number of uniformly excited aperiodic concentric rings to match a predefined reference pattern. An auxiliary layout of concentric rings based on the Bessel function approximation of the circular array pattern is first synthesized using the off-grid CS framework and convex optimization to match the reference pattern using a minimum number of rings. Then, a density tapering technique is used to replace each ring with a ring of uniformly spaced isotropic radiating elements. Simulations were carried out to assess the effectiveness of the proposed method. The final part of the thesis proposes an algorithm based on group-sparse recovery and convex optimization for thinning multi-carrier frequency diverse arrays (FDAs). The algorithm ensures that the turned off antennas have a negligible effect on the beampattern for all multi-carrier frequencies assigned to them. Numerical results showed that the proposed algorithm can save about 25% of the antenna elements compared to the existing synthesis methods under the required constraints on the radiation pattern.