High-Dimensional Uncertainty Quantification for Radar Cross Section of Multi-Layered Dieletric Block

We propose an uncertainty quantification method for Radar Cross Section (RCS) of multi-layered dieletric block. Due to the numerous components and complex shape of the dielectric block, traditional intrusive uncertainty quantification methods, requiring significant modifications to the origin solver, are no longer feasible. In addition, the important material parameters of each multi-layered dielectric block component are considered in uncertainty quantification, leading to high-dimensional uncertainty. As a result, the uncertainty quantification of RCS becomes even more challenging. To address these issues, based on a data-driven ideology, we construct a surrogate model from a sparse dataset to enable fast prediction of the RCS, and thereby extract the statistical characteristics of the RCS to achieve rapid and efficient uncertainty quantification.