摘要
针对多约束条件下阵列优化过程中运算复杂和波束指向固定的问题,提出一种基于混合整数有向差分进化(mixed integer directed differential evolution, MIDDE)算法的综合方法。首先,通过几何变换减少优化目标数量,缩小寻优空间,降低运算复杂度;其次,对差分进化(differential evolution, DE)算法的变异过程引入随机选择与排序策略,提高搜索速度与精度,并改进边界约束处理方式,提高边界搜索能力。最后,用MIDDE算法优化基于正则化稳健超指向波束形成器的目标函数。仿真结果表明,在白噪声增益(white noise gain,WNG)受约束的情况下,优化阵列可在宽波束指向角范围内获得较高的方向性因子(directivity factor,DF),且显著高于相同阵元数的规则阵列,该方法可有效提高阵列性能。
Aiming at the problems of complex operation and fixed beam direction in array optimization under multiple constraints,a synthesis method based on mixed integer directed differential evolution(MIDDE)algorithm was proposed.First,the number of optimization targets was reduced,the optimization space was reduced,and the computational complexity is reduced through geometric transformation;Secondly,random selection and sorting strategy were introduced into the mutation process of differential evolution(DE)algorithm to improve the search speed and accuracy,and the boundary constraint processing method was improved to improve the boundary search ability.Finally,MIDDE algorithm was used to optimize the objective function of the regularized robust super directed beamformer.The simulation results show that under the constraint of white noise gain(WNG),the optimized array can obtain a higher directivity factor(DF)in a wide beam pointing angle range,which is significantly higher than the regular array with the same number of elements.This method can effectively improve the array performance.
作者
张正文
汪震
廖桂生
ZHANG Zheng-wen;WANG Zhen;LIAO Gui-sheng(School of Electrical and Electronic Engineering,Hubei University of Technology,Wuhan 430068,China;State Key Laboratory of Radar Signal Processing,Xi’an University of Electronic Science and Technology,Xi’an 710071,China)
出处
《科学技术与工程》
北大核心
2023年第7期2892-2900,共9页
Science Technology and Engineering
基金
湖北省自然科学基金(2018CFB545)
中国博士后科学基金(2016M600729)。
关键词
麦克风阵列
差分进化算法
阵列几何优化
MINLP
microphone array
differential evolution algorithm
array geometry optimization
MINLP
