Research on Radar Cross Section Scaling Relation based on Neural Network

Radar Cross Section (RCS) is an important indicator of stealth aircraft. In the stage of design and engineering development, RCS measurements are often performed on scaled models to verify design performance. For absorbing materials and composite materials, it is difficult to meet the constraints of the similarity principle. The difficulty lies in the difficulty of finding materials that meet the constraints of electromagnetic parameters. By establishing a BP neural network model for the scale prediction problem of the surface-coated absorbing material model, and using the network to predict the data under different polarizations, an improved scheme is proposed to optimize the prediction performance in the low scattering area. The results show that the improved BP artificial neural network model can reduce the estimated average error to 1dB, and the performance is good.