Grating Lobe Suppression of Non-Uniform Arrays Based on Position Gradient and Sigmoid Function

• Main Authors: Xiaomin Xu, Cheng Liao, Liang Zhou, Fan Peng
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Grating lobe suppression; gradient; large non-uniform arrays; optimization algorithm; sigmoid function
• Online Access: https://ieeexplore.ieee.org/document/8782093/

A super-multivariate optimization algorithm is proposed to suppress the grating lobes (GLs) of non-uniform arrays. For position-only variable cases of any huge array, it is always difficult to deal with by using clustering algorithm because of thousands of variables. The proposed method achieves grating lobe suppression (GLS) by taking position gradient of dynamic maximum GL and obtaining phase difference by comparing with every element, which requires only precious few extra computations. When gradient function varying with each element position is established, an appropriate sigmoid function will be introduced and used to control the magnitude of small element displacement. After each generation of element-movement operation, electric field of all elements are superimposed and partially cancelled out each other in the target direction. In addition, the gradient computing does not significantly increase the cost of total computation under multi-variable conditions, which avoids the unacceptable cost of conventional clustering optimization algorithm. In this paper, a 16×16-element array is optimized, and a result comparison with random optimization (RO), particle swarm optimization (PSO), and gradient algorithm (GA) is present. The validity of the proposed algorithm for GLS is verified through using a 100×100-element array. In the final, a large array antenna based on subarrays is optimized for a low sidelobe level (SLL) below -20dB using the proposed algorithm. The results meet the requirements and show the effectiveness of the proposed algorithm.