Inverse Artificial Neural Network for Multiobjective Antenna Design

• Published in: IEEE transactions on antennas and propagation 2021-10, Vol.69 (10), p.6651-6659
• Main Authors: Xiao, Li-Ye, Shao, Wei, Jin, Fu-Long, Wang, Bing-Zhong, Liu, Qing Huo
• Format: Article
• Language: English
• Subjects: Algorithms; Antenna design; Antenna radiation patterns; Antennas; Artificial neural networks; Computational modeling; Integrated circuit modeling; inverse artificial neural network (ANN); Learning theory; Machine learning; Microwave antennas; multiobjective; Neural networks; Optimization; Performance indices; Training; transfer function (TF)
• ISSN: 0018-926X; 1558-2221
• DOI: 10.1109/TAP.2021.3069543

To improve the convenience and efficiency of antenna design, in this article, a novel inverse artificial neural network (ANN) model is proposed in which antenna performance indexes are set as the input and corresponding geometrical variables are set as the output. To solve the multiobjective problem of antenna modeling, the first part of the ANN model involves three parallel and independent branches for S -parameter, gain, and radiation pattern, and the second part outputs a final predicted result. Once the training is completed, the proposed inverse model can provide antenna geometries directly without being repetitively called by an optimization algorithm. Compared with the inverse extreme learning machine and five state-of-art forward models, the proposed model uses a small number of training samples and directly obtains the satisfying values of geometrical variables without any optimization process.