Synthesis Of Large Antenna Arrays With Sparse Reconstruction

Due to the advantages of high gain,narrow beam and strong ability of beam steering,the large array antennas have been applied more and more extensively in the modern radar and wireless communication systems.Meanwhile,in order to reduce the cost and system complexity,the large array antennas are expected to be designed with sparse arrays or subarrays.However,the existing design methods still face some challenges.The thesis studies several key technologies in the design of large array from the angle of practical application including sparse array synthesis,design of large-spaced array and sparse array with pattern reconfigurable antenna and subarray technique.Furthermore,this thesis emphatically studies on the application of sparse reconstruction in the synthesis of large array antennas.The main works of the thesis includes the following five parts: 1.Synthesis of sparse array using compressive sensingThis thesis examines the problem of maximally sparse array synthesis from the point of view of sparse signal processing.A CS model for the synthesis of sparse array is established where the problem of maximally sparse array synthesis is converted into a problem of sparse reconstruction with linear constraint.In the framework of CS theory,a method based on the FOCal underdetermined system solver(FOCUSS)is proposed for the design of maximally sparse array.This method can intelligently determine the minimum number of required elements,element locations and excitations,and it is applicable for arrays of arbitrary topological type,including linear array,planar array and conformal array.2.Design of multiple-pattern sparse array using MMVCSRA new technique based on the multiple measurement vectors FOCUSS(M-FOCUSS)for the synthesis of pattern reconfigurable sparse arrays is presented.The method,based on the multiple measurement vectors collaborative sparse recovery(MMVCSR)theory,is used for finding the common element positions and individual element excitations for multiple patterns in the guidance of collaborative synthesis strategy.Then,this method is combined with the active element pattern(AEP)technique so that the mutual coupling between real antennas can be taken into account in the synthesis process.Simulated and measured results have validated the effectiveness of the proposed method.3.Fast synthesis of large array using perturbed compressive samplingFirstly,an effective method based on the perturbed compressive sampling(PCS)is proposed for the sparse array synthesis.Position perturbation variables are augmented to the traditional CS-based model,which allows continuous element placement.Such a strategy extends the optimization space of sparse array synthesis problem and reduces the required number of grids for the algorithm which significantly improves the optimization effects and computational efficiency.Then,in order to further reduce the modelling error and the computational cost,an extended PCS(EPCS)model with a secondary grid strategy is presented.Thirdly,in order to overcome the problem that the element positions will have a imaginary component when dealing with complex-valued cases,an alternating iterative algorithm which alternately iterates between a sparse recovery process and a local optimization procedure is presented and applied to the synthesis of complex-excitation and pattern reconfigurable sparse arrays.Finally,the paper studies the application of the proposed PCS-based method in the synthesis of large sparse and steerable arrays with low sidelobe level(SLL).4.Design of large-spaced array sparse array with pattern reconfigurable antennaFirstly,a pattern reconfigurable antenna(PRA)which can achieve two-dimensional and circular beam switching.Then,the PRA is applied to the design of sevral kinds of sparse arrays including rectangular grid large-spaced array,triangle grid large-spaced array and random sparse array.Compared with the traditional phased array,the proposed arrays have distinct advantages in terms of array sparsity,scanning range,scanning gain and SLL.5.Grating lobe suppression with regular subarray architecturesA technique for reducing grating lobe in phased arrays with regular subarray architectures is proposed.The method is based on the subarray modules with different beam directions and the switching scheme of the subarray modules.The grating lobe can be effectively suppressed through optimizing the beam direction of each subarray pattern and selectively excite the subarray modules during the scanning process.The regular subarray is equivalent to a nonuniform sparse subarray whose subarray modules have different beam direction and the architecture will change with the scan angle.By using this approach,the regular subarray can achieve the similar 3d B beamwidth with the element-level phased array having the same aperture.The reduction of the channels is larger than the gain loss.