Research On Synthesis Method Of Beam Scanning In Sparse Array Antenna

As an important technology in the fields of radar engineering,sparse array synthesis technology has drawn many researches'attention.Compared with conventional uniformly distributed full array antennas,sparse arrays can produce the same radiation beam width,similar antenna gain and lower sidelobe level,but their cost and power consumption are much lower than traditional uniformly distributed full array antenna's.Due to the high complexity and the large amount of calculation of sparse array synthesis technology,most of the related researches focused on sparse arrays with fixed beam pointing while there are few reports on the synthesis of sparse arrays with wide-angle scanning beams.The genetic algorithm can obtain the global optimal result,but for the multi-dimensional multi-parameter optimization problem,it needs great amount of calculation so it is rarely applied to beam scanning sparse array synthesis.The Gaussian process regression algorithm makes predictions based on existing samples and can be used to improve optimization efficiency.Therefore,based on the genetic algorithm,the MEMP algorithm and the Gaussian process regression algorithm,this paper studies the synthesis method of sparse arrays to reduces the calculation amount of the comprehensive process,and improves the optimization efficiency.The main tasks are as follows:First,the theory and implementation scheme of the proposed method are introduced in detail.This method solves the aperture constraint of the sparse array and the minimum array element spacing,and nests a two-layer genetic algorithm to optimize the array element position and the corresponding excitation distribution.Aiming at the problem of over-computation caused by the optimization of complex nonlinear problems with high latitude and multi-parameters,this comprehensive method incorporates augmented matrix beam algorithm to provide initial values for real-number genetic algorithms,which improves the convergence speed of genetic algorithms At the same time,the Gaussian process regression algorithm is added to the comprehensive method to replace part of the calculation with prediction,which can effectively reduce the overall amount of calculation.Secondly,using the proposed synthesis method,three sparse linear arrays at the aperture of15.5?₀,11.5?₀ and 9.5?₀are designed for different types of antenna radiating elements,and the corresponding uniform full arrays are 1×32,1×24,1×20.The resulting sparse array has 19,15,and13 units,respectively,reaching sparse rates of 40.625%,37.5%,and 35%.The comprehensive results are compared with the results of the full-wave analysis to verify the effectiveness of the proposed algorithm in designing sparse linear arrays.Compared with the uniform full array,the HFSS simulation results of sparse arrays composed of the three units show that the sparse array can achieve a beam scanning effect with a similar beam width,lower side lobe level.Due to the limitation of the sparse rate and the number of iterations,the scanning range of the sparse linear array beam scanning becomes narrower.Finally,using the proposed synthesis algorithm to synthesize the planar array,the beam scanning sparse synthesis is carried out for the two different types of antenna elements at the aperture of4.5?₀×4.5?₀ and 5.5?₀×5.5?₀,and the size of the corresponding uniform array is 10×10 and12×12 separately.The sparse arrays synthesized have 49 and 81 units respectively,reaching a sparse rate of 51%and 43.75%.The sparse array composed of the test unit 1 has a radiation characteristic close to a uniform full array in the range of±20°,and the sparse array composed of the test unit 2can achieve a beam scanning effect similar to the full array in the range of±40°.The HFSS full-wave simulation was used to verify the above sparse array design results.The results show that the characteristics of the radiation main lobe are in good agreement.Since the mutual coupling between the elements is not considered,the characteristics of the radiation side lobe of the two are poor.This work provides a useful reference for the next step in the design of large-scale sparse area arrays.