Research On Advanced Technique Of Array Antenna Pattern Synthesis

Antenna array has been widely used in communication,radar,navigation and satellite fields,due to its predefined radiation characteristics.As one of the core technologies of array antennas,array pattern synthesis refers to by optimizing the amplitude and phase weights,the number of array elements,and the positions of the array elements according to the predefined pattern characteristics,near ideal array pattern performance can be achieved With the continuous development and advancement of technology,the performance requirements raised by the system on antenna array are increasing,such as,low sidelobe high gain and flexible control on radiation pattern.As an important method of optimization,the capability of the array synthesis algorithm determines whether the array antenna can achieve the desired radiation pattern or not.At present,the existing array antenna pattern optimization algorithms show several shortcomings,such as narrow application,high computational complexity,slow convergence,too many control parameters,difficult debugging,inaccurate optimization results,and inappropriate to apply to actual scenarios.Therefore,it is of great significance and application value to explore new technology for array pattern synthesis with wide application range,fast convergence speed,easy implementation and stable results.This paper mainly focuses on the research work of convex optimization based array synthesis algorithm,neural network based array synthesis algorithm,and conformal array pattern synthesis.The research content is mainly divided into three parts:First,when applied to beam forming and amplitude excitation constraint,the existing convex optimization algorithm has some defects,such as reduced solving area,increased complexity and many control parameters.To address the above challenges,a new static and dynamic convex optimization algorithm is proposed.The proposed algorithm can effectively expand the solution area without increasing the control parameters and complexity.In addition,shrinkage factor and crossover operator are introduced into the dynamic convex optimization to ensure the convergence and stability.The proposed algorithm is verified by appling to beamforming,phase-only pattern synthesis and reconfigurable array pattern synthesis problems,and is compared with existing advanced optimization algorithms.The results show that the proposed algorithm has the advantages of better radiation characteristics,faster convergence and more stable results.Second,in view of the shortcoming that the increase in the number of array elements leads to a sharp drop in the optimization speed,inspired by the powerful search ability of neural network,an array optimization algorithm based on encoder and decoder is proposed.Wherein,the encoder optimizes the excitation while the decoder calculates the excitation pattern.When the output results satisfy the target pattern,the encoder stops optimizing the excitation.In order to make the same decoder can deal with different array numbers,virtual array elements are introduced into the data set of the decoder and set them to zero randomly.In the simulation experiment,training results of decoders with different structures are first verified,which show that the algorithm can obtain suitable decoder structures.Then,the same decoder is used to optimize the arrays with different number of elements and spaces.The results show that the algorithm can maintain good optimization results and converge very fast.Third,the two optimization algorithms mentioned above are applied to the array synthesis of the conformal antenna array.aiming at the challenge that the orientation pattern of conformal antenna array is difficult to be calculated due to the different orientation of elements,a rational selection of active element pattern for calculation is proposed.Next,considering the importance of the polarization mode in the conformal array,polarization constraints are introduced into the algorithm to ensure that the polarization mode satisfies the requirements.Finally,the proposed optimization algorithm is applied to the array synthesis of cylindrical conformal arrays.Simulation results show that both of the two methods can complete the excitation optimization within quite short time,and the obtained patterns can satisfy the benchmark requirements,thus verifying the effectiveness of the proposed methods.