Adaptive sampling method for experiments with fixed-grid variables in measurements of nonlinear active devices

This paper describes an adaptive sampling method suitable for experiments where values of at least one of the input variables are limited to a fixed set. The method is evaluated on amplifier measurements using a calibrated nonlinear vector network analyzer. The input variables were the input power, varied in a continuous way, and the fundamental frequency, limited to sparse and dense sets of fixed values. The adaptive sampling algorithm enabled to sample more densely in the highly nonlinear regions, and evenly in the linear regions.