A Novel Intelligent Antenna Synthesis System Using Hybrid Machine Learning Algorithms

Abstract

The synthesis of antenna is crucial to the development of wireless communication. Traditional synthesis method is complex, time consuming and requiring the designer being familiar with the antenna topology and techniques. With the development of artificial intelligence technology, the efficiency and accuracy of antenna synthesis can be improved with machine learning algorithms. In this paper, a novel hybrid machine learning model is proposed to synthesize antenna based on the performance requirements of the antenna, which requires a small number of training dataset and has higher accuracy than a single machine learning model. In this model, geometrical parameters analysis is first performed to select the parameters that have the greatest impact, which requires fewer samples for training. After parameters selection, we propose a new hybrid machine learning model to map performance characteristics onto the geometrical parameters. This model combines 10 different machine learning models (base learners) to improve the generalization performance of the entire model by learning the advantages of their respective base learner. The hybrid model possesses the best prediction ability with a satisfied mean square error (MSE) of 0.00456. Finally, the triangular pin-fed patch antenna is used to demonstrate the validity and efficiency of this proposed model.