Research On The Synthesis Techniques For Sparse Antenna Arrays With Low Sidelobe Level

The synthesis technique for aperiodic arrays has aroused great interest in array synthesis.Though the array gain of the aperiodic array has been reduced,compared to the uniformly spaced one with the same array aperture,the main beam width has kept almost the same width and the directivity is still strong,meanwhile,the complexity of the antenna system,weight and cost can be greatly reduced.Therefore,the aperiodic array has been widely studied in the radar,satellite communication systems,and radio astronomy.Aperiodic arrays are categorized into non-uniform and uniform amplitude ones,compared to the former,the later can further simplify the feeding networks,degrade the complexity and cost,and be overwhelmed with manufacture and maintenance.Unfortunately,sidelobes of the aperiodic array pattern are changed with element positions severely,and the peak-sidelobe level(PSLL)is often at a high level,thus the element positions must be optimized so as to reduce the sidelobe level.The optimization of element positions is indeed to solve a complicate nonlinear problem,and massive optimum algorithms are inefficient due to the huge computational complexity of array patterns,in particular,effective techniques are absent for large scale sparse arrays from the current reports.This dissertation presents several new synthesis techniques to the sparse linear and planar arrays,in which the thinned array that the element positions can be only selected from the regular lattice and the sparse array within or without the internal element spacing constraint are included.Note that a minimum element spacing constraint sparse array is called when the minimum value of all the element spacings is constraint to larger than a preset const.In terms of uniform amplitude linear arrays,the relationship of Fourier transformation between array factor and element positions of the thinned array is discussed,as the element positions are distributed arbitrarily,the equalization relationship between array factor and frequency response of an FIR filter is studied,besides,the affection of sidelobe level that suppressed by weighting density sampling with a minimum element spacing constraint is qualitatively analyzed.Three related deterministic techniques are studied,i.e.a modified iterative Fourier technique(MIFT),and approaches based on equivalent formulation and weighting density respectively.Furthermore,a new two dimensional conception of weightingdensity extended from the linear array is built to optimize planar arrays with a minimum element spacing constraint,as the synthesis procedure of concentric ring arrays and thinning strategy of square or triangle grid arrays are involved.The deterministic techniques are featured with low computational complexity,but the achieved results are usually demonstrated to be inferior ones compared to randomly search algorithms.In order to reduce the computational complexity of randomly search algorithms that paid during the optimization of small scale arrays,the integer margin code and nonlinear mapping modeling are studied.In view of the potential performance,practicability,redesign costs and flexibility,genetic algorithm is selected as a revisable method,and an improved integer genetic algorithm(IIGA)is developed for searching the global optimum results of small-scale sparse arrays.In particular,a totally new recombination of margin code technique(RMCT)aiming at further reduction of computational complexity in the search algorithms is studied,with respect to the optimization of one-dimensional element positions.Since the solution space of element positions is exponentially extended along with the increased element number,applying a randomly search algorithm directly cannot suppress the PSLL of a sparse array down to a satisfied level within a limited time span.Through qualitative analysis of the intrinsic difficulty existed in the optimization of large scale sparse arrays,this dissertation studies an innovative strategy as well as a hybrid synthesis modeling to constrain the optimum solution space for sparse arrays with the minimum element spacing constraint.A closed form of continuous weighting density is derived,besides,the relationship between ring radii,number of elements on a ring,continuous weighting density,and the total element number of a reference array is achieved when the hybrid synthesis modeling applied to concentric ring arrays or square grid arrays with circular separation.Consequently,based on the hybrid synthesis modeling,three different optimization techniques that aiming to different array synthesis circumstances are proposed separately,which provide alternative solutions to the optimization problems of large scale sparse arrays.