ADMM-based array antenna synthesis algorithm with reconfigurable pattern element selection

Abstract

The paper considers the array beam pattern synthesis (BPS) problem with reconfigurable pattern elements (RPE). Both the array weight and the array pattern are optimized, which further suppresses the peak sidelobe level (PSLL) compared with the fixed pattern array. For the array pattern optimization, the conventional selecting of appropriate N elements from K radiation patterns requires to evaluate all possible KN choices, resulting in the non-deterministic polynomial-hard RPE selection problem. In this work, this problem is solved through investigating different element pattern combinations with randomly arranged pattern locations. In such way, the possible solutions to the RPE selection problem has an overall choices of CN+K−1K−1=(N+K−1)!N!(K−1)!, which extremely shrinks the solution space. An algorithm with the RPE selection and array weight optimization is proposed for the BPS problem. Furthermore, the alterative direction method of multipliers (ADMM) framework is introduced to accelerate the computation speed. The proposed algorithm provides an effective scheme for generating appropriate RPE combinations for PSLL suppression, resulting in a wider-angular scanning ability for array antenna. Simulation results demonstrate 0 dB to 4 dB lower PSLL can be obtained compared with the algorithms based on the fixed element pattern.