Plasma Parameter Inversion Base on Deep Learning Approach

Abstract

As a dispersive medium, plasma has broad application prospects in stealth antenna and new attenuator design. Plasma parameters, such as electron density and collision frequency, are the basis for studying plasma property. At present, conventional plasma parameter inversion algorithms encounter some difficulties, such as large amount of calculation, long inversion time, low inversion accuracy. A deep learning model for plasma parameter inversion is proposed in this paper. The model takes full advantage of the characteristics of deep neural network, with simple structure, fast operation speed. And the plasma parameters can be reconstructed with high precision. The simulation results indicate that the inversion results are better than traditional methods even in the presence of noise.