Adaptive sparse grid method for RCS estimation of object with uncertain shape

Presented is an adaptive sparse grid method to efficiently estimate the radar cross section (RCS) of a 2D cylinder with uncertain holes, as well as uncertain locations, sizes and number of the holes. This model has been applied extensively in microwave engineering, e.g. frequency selective surface (FSS). Stochastic collocation method is employed to solve the resulting stochastic differential equation. The Gauss-Patterson formula is used in the sparse grid method. The adaptive sparse grid method is based on Richardson extrapolation. Numerical results show the proposed method works very well.