A Two-dimensional Surrogate Model Technique for Electromagnetic Scattering Analysis of 3-D Objects with Varying Material and Shape

A two-dimensional surrogate model technique of 3-D objects with varying material and shape is pretended. The technique based on gaussian process regression (GPR) constructs surrogate models to predict the monostatic RCS of target. This method takes the relative permittivity and geometric size of the 3-D object as the training inputs, while utilizes the RCS values as the training outputs. A series of training datasets are constructed using the principles of sampling. Through a series of test inputs involving variations in material and shape of the 3-D object, the prediction results of RCS obtained from the surrogate model demonstrate high accuracy and improved efficiency in electromagnetic scattering analysis.