Machine Learning Based Inversion Method for 3-D Plasma Parameters

This paper applies the machine learning method to plasma parameter inversion. Firstly, we calculate the far-field scattering field of a 3-D metallic sphere coated with homogeneous magnetized plasma using the current density convolution finite-difference time-domain (JEC-FDTD) method. By varying plasma parameters in the forward model, we obtain a series of scattering field data to create a dataset for training and testing. Machine learning models are then designed to reconstruct the plasma parameter, including the support vector regression model, the fully connected artificial neural network model, and the 1-D convolutional neural network (1D-CNN) model. The experimental results show that the 1D-CNN model has higher accuracy and superior performance.