Multivariate Look-Up Table Based on N-Linear Interpolation for General Reflectarray Design

Abstract

We present a multivariate look-up table (LUT) based on N-linear interpolation for a fast general reflectarray design. In order to achieve an efficient implementation of the LUT, details are provided on its structure and a technique for a fast memory access of the reflection coefficients is described. A general formulation for the N-linear interpolation is given and then it is applied to a general reflectarray design procedure, including the improvement of the cross-polarization figures of merit. In order to benchmark the computational efficiency of the LUT, it is compared with other two tools, a method of moments based on local periodicity and surrogate models based on support vector regression. Results show that the LUT achieves a high computational efficiency while preserving a high degree of accuracy.