Research On Subarray Array And Base Station Antenna Based On Improved Fruit Fly Algorithm

With the rapid development of antenna technology,the design requirements for antennas are getting higher and higher.Compared with the original antenna form,the antenna element has developed from a monopole antenna to microstrip antenna,and the antenna array has been developed from an initial linear array to the planar array,the spherical conformal array,the thinned array and other array methods.However,with the development of antenna form and antenna array,the design difficulty is also increasing.Therefore,regardless of the antenna element or the antenna array,new design ideas and design methods are required.In phased array antenna systems,most of them increase the electromagnetic performance of the entire antenna system by increasing the number of antenna elements to form a large linear array or planar array.However,the full array design brings great challenges to the array feed system design,including the inability to guarantee the amplitude consistency of the array elements,the overheating of the antenna system caused by too many TR components,the need for additional cooling,and the high cost caused by the above problems.Therefore,in continuous research,antenna designers begin to divide several antenna elements into the same group according to a certain rule,and use the same amplitude or phase excitation to form a sub-array.The total number of sub-arrays will be less than the total number of antenna elements,greatly reducing the design cost.However,due to the common amplitude and weighting of the elements in the sub-array,when the pattern of the sub-array is synthesized,the degree of freedom is reduced and the antenna array fails to meet the index requirements.Therefore,an array antenna pattern synthesis method with excellent performance must be proposed.The antenna array is composed of antenna elements,so the electromagnetic characteristics of the antenna elements will also greatly affect the radiation characteristics of the antenna array.When solving the problem of antenna element size optimization,it is necessary to use the electromagnetic simulation software for simulation,and if full-wave electromagnetic simulation is performed for each parameter,huge calculation cost and time cost will be consumed.Existing optimization methods including intelligent optimization algorithms and electromagnetic simulation software joint simulation methods or multi-threaded parallel operation simulation use CPU or GPU to reduce design time.However,the realization of the above two methods requires the help of powerful computer performance,and does not really reduce the design cost.This thesis combines the three aspects of subarray array antenna pattern synthesis,new antenna size optimization methods and base station subarray antenna beamforming.The research work is mainly divided into three parts:Firstly,an improved fruit fly algorithm based on particle behavior mechanism,orthogonal experiment mechanism,and simulated annealing mechanism is proposed to improve the diversity of the population while avoiding falling into local optimality,which improves the optimization performance of the algorithm.At the same time,the performance of the algorithm is analyzed and compared with a number of improved fruit fly algorithms.When the optimization range is limited or the selection of the extreme point is changed,the superiority of the algorithm performance is fully demonstrated.Comparing the improved algorithm with other kinds of algorithms,the algorithm still achieves more significant optimization results.This shows that the improved fruit fly algorithm is more suitable for engineering applications.Then,the algorithm is applied to solve the synthesis problem of the subarray array pattern,and the sidelobes of the pattern are optimized in the application examples of the four subarray arrays.The performance is compared with other optimization methods.Although the complexity of the subarray-level array antenna is much more complex than that of the ordinary antenna array pattern synthesis,compared with other algorithms,the subarray-level array antenna pattern synthesis method based on the improved fruit fly algorithm can still be quickly and effectively obtained.The optimization results prove the effectiveness of the proposed method in synthesizing the subarray antenna patterns.Finally,for the optimization of antenna size,an antenna optimization method based on xgboost machine learning algorithm and the improved fruit fly algorithm is proposed.First,a rough data set is established for the design parameters,and then training is performed through xgboost.The sensitivity of each parameter to the radiation characteristics is analyzed.For further training,a data set is established for the highly sensitive parameters.During the training process,the improved fruit fly algorithm is used to optimize the model parameters to obtain the machine model with the best prediction effect.Then take the antenna design requirments as the optimization target and the antenna design parameters as the optimization variables,search and optimize in the obtained prediction model,in order to obtain the final optimized size parameter.This method is applied to the miniaturization and broadbandization of base station antennas.Experimental results show that this method reduces the design cost of the antenna design while ensuring design accuracy.At the same time,the base station antenna is used to form a subarray-level antenna array and beamformed.