阵列天线方向图修正算法  

Correction Algorithm of Array Antenna Patterns

作　　者：张鑫 万明[2] 陆德江 Zhang Xin;Wan Ming;Lu Dejiang(Graduate School,Air Force Engineering University,Xi’an 710038,Shaanxi,China;Aviation Engineering School,Air Force Engineering University,Xi’an 710038,Shaanxi,China)

机构地区：[1]空军工程大学研究生院,陕西西安710038 [2]空军工程大学航空工程学院,陕西西安710038

出　　处：《激光与光电子学进展》2022年第17期138-146,共9页Laser & Optoelectronics Progress

摘　　要：为进一步优化天线阵列激励得到的宽零陷和低旁瓣的方向图,通过软件Matlab2019b对天线阵列波束成形算法进行仿真,对比方向图分析说明,在原有约束的基础上增加二次约束以及改进协方差的方式存在展宽主瓣的问题。为了在展宽零陷和抑制旁瓣的同时保持主瓣宽度不变,提出了基于幅值十分位粒子群并行寻优的线性约束最小方差(LCMV)方向图修正算法。该算法通过对比几种算法所得到的天线阵列激励的特点,引入先验信息,仅对权矢量的幅值寻优缩小可行解空间,同时改进粒子群算法寻优机制,通过十分位寻优方法使算法收敛更稳定,并且对粒子群算法代码向量化,同时利用图形处理器使每个粒子同时更新,实现粒子群并行算法,加快算法计算时间。仿真结果表明:该算法能实现展宽零陷和低旁瓣的同时,保持主瓣宽度,在所对比的算法中效果最好,同时基于幅值十分位粒子群并行算法收敛所需的迭代次数更少,计算速度更快,且天线规模越大提升越明显。Antenna array beamforming algorithm is simulated using Matlab2019b to optimize antenna array excitation for obtaining wide nulls and low sidelobe patterns.Comparative pattern analysis shows that adding quadratic constraints based on the original constraints and improving the covariance have the drawback of widening the main lobe.This study proposes a linear constrained minimum variance(LCMV)pattern correction algorithm based on parallel optimization of amplitude decile particle swarms to maintain the width of the main lobe while maintaining wide nulls and low sidelobes.The algorithm compares the characteristics of the antenna array excitation obtained using several algorithms,introduces preliminary information,optimizes only the magnitude of the weight vector to narrow the feasible solution space,improves the optimization mechanism of the particle swarm algorithm,and adopts the decile optimization method.The algorithm convergence is made more stable,and the particle swarm algorithm code is vectorized.A graphics processing unit is used to simultaneously update each particle,realize the particle swarm parallel algorithm,and speed up the computational time of the algorithm.The simulation results show that the algorithm can achieve a widening of nulls and low sidelobes while maintaining the width of the main lobe.In addition,it has the best effect among the compared algorithms.Simultaneously,the convergence of the particle swarm parallel algorithm based on the amplitude decile requires fewer iterations and it has faster computational speed,and the improvement is more obvious when the antenna scale is larger.

关 键 词：傅里叶光学与信号处理 波束形成 粒子群优化算法 并行计算 零陷展宽 

分 类 号：TN991.[电子电信—信号与信息处理]