Research On Pattern Synthesis Of Large-scale Sparse Array And Subarray Technology

Pattern synthesis technology has always been the core problem,difficult problem and hot issue in the field of array signal processing.This is a technology that controls the power gain of the receiving or transmitting beam in different spatial angles by controlling the weights of different array elements.It is widely used in modern radars,communication systems,sonar and other fields.In recent years,the scale of the array has developed in the direction of large-scale.While obtaining good performance,the high computational complexity and complex feeder network after the large-scale array make the system implementation extremely complicated.Aiming at some of the problems of large-scale arrays,this paper is based on practical application requirements,focusing on reducing the computational complexity and design complexity of large-scale arrays,and proposes k-means algorithm based on k-means algorithm combined with convex optimization and genetic algorithm(Genetic Algorithm(GA)hybrid method divides the subarray technology and proposes a sparse array design technology.The research content is summarized as follows:At first,a subarray technology based on pattern synthesis of large linear and planar arrays is improved.Using the relationship between subarray division and optimal weights,the problem of subarray division based on pattern synthesis that approximates the desired pattern is simplified to the problem of determining the optimal sub-array configuration(subarray layout and corresponding sub-array weights),Further transform the problem into a clustering problem,and then use the K-means clustering algorithm to solve the problem.The pattern obtained by the proposed algorithm can obtain a lower peak side lobe level(Peak Side Lobe Level,PSL)while maintaining a narrow main lobe width.Next,by combining convex optimization and GA,a hybrid method of dividing subarrays is improved to solve different optimization problems(optimum and beam problems,low PSL pattern synthesis and approximation to the reference pattern problem)for the optimization of uniform linear arrays Subarray layout and corresponding optimal subarray weights.The optimization problem of dividing the subarray is nonlinear and non-convex,which is difficult to solve directly.The non-convex problem can be solved by using GA.By making the individuals in the GA population correspond to the subarray division scheme,for the determined subarray layout,the above-mentioned optimization problem It degenerates into a convex optimization problem only about sub-array weight variables.The convex optimization problem about sub-array weight variables is used as the fitness function of GA,and the corresponding optimization parameter is the fitness value.Then use the convex optimization to calculate the fitness value of the individual and calculate the optimal subarray weight.At last,a sparse array design algorithm based on iterative weighting for constant beamwidth of broadband signals is improved.The norm is approximately replaced by the norm.In each iteration of the calculation process,equal weights are applied to a group of taps of the same array element.The weight value of the same element is controlled to change toward the same trend.After multiple iterations,the norm of the tap weight vector is much smaller than other elements.The element is considered to be inactivated,reaching sparseness.The purpose of the design.