Optimization design of skin antenna based on Bayesian optimization

Abstract

Skin antenna consists of microstrip antennas, skin and honeycomb, and it is widely used in wireless applications of air, water and vehicles. Skin antenna is not only a load-bearing skin structure, but also a microwave antenna which can receive or send electromagnetic waves, and the skin and honeycomb were designed to bear the force. Thus the thickness of them has a large impact on the capacity of the bearing. In order to quickly determine the size of the antenna element and the thickness of skin and honeycomb, this paper presents a fast optimization method for the electromechanical co-design of skin antennas. First, the length and width of the antenna element are taken as the optimization variable, and the return loss is used as the optimization target when the antenna unit is optimized. Then the gain of the antenna array is used as the optimization target by optimizing the thickness of each layer of skin and honeycombs to realize the optimization design of skin antenna array. The return loss and the upper and lower boundary of the design variable range are taken as constraint conditions. The final size of the skin and the honeycomb is determined by using Bayesian surrogate model. Finally, the simulation results are compared with the measured results, and experiment results have revealed the feasibility and effectiveness of the optimization method for skin antenna design.